# Agent Source: https://docs.trymeridian.com/agent Ask questions about your Meridian data and get a clear explanation plus recommended next steps. Agent helps you move from “something changed” to “here’s what to do next.” It is most useful when you want a fast explanation of a KPI shift, have questions about the Meridian product, or when you know what outcome you want (more visibility, better rank, more citations) but you are not sure which lever to pull first. Agent Agent uses your tracked prompts, recent runs, citation drivers, and your configured competitors/products to recommend actions inside Meridian. Agent is most helpful after you have added a meaningful prompt set and competitor set. If you only track a few prompts, the recommendations can be correct but less specific. ## What Agent is good for ### Explaining changes Agent can explain changes on Home, including: * why visibility dropped, * why citation rate changed, * why prominence got worse, * and which sentiment dimension is trending negative. ### Prioritizing what to do next Agent can recommend the most impactful next steps, like: * which prompts to investigate first, * which sources to target for outreach, * which pages to fix on your site, * which opportunities to publish next. ### Answering “how do I” questions Agent can help new users learn Meridian quickly, for example: * “How do I diagnose a prompt?” * “How do I improve citations?” * “How should I use Off-page?” ## Example questions that work well Here are prompts that usually lead to helpful, actionable answers: * “What changed this week on Home, and what should we do about it?” * “Which prompts are we most often not mentioned in, and what do those winners cite?” * “Which citations drivers should we prioritize for outreach?” * “Which website pages should we fix first to improve citations?” # Citations Source: https://docs.trymeridian.com/analytics/citations Identify the sources AI trusts in your category and build a plan to get included. Citations show the source ecosystem behind AI answers. If you are missing from the sources AI relies on, you will often struggle to rank consistently—especially on “best X” prompts where trust and evidence matter. Citations dashboard showing source breakdown and drivers ## Left vs. right panels (what each one means) The Citations page is split into two panels because there are really two different questions you need to answer: * ***“What sources does AI use when it talks about us?”*** (left) * ***“What sources does AI use when it answers the prompts we care about?”*** (right) Understanding the difference is important because the actions are different. The left panel tells you how well your own content is serving as “the source of truth” about your brand. The right panel tells you which parts of the internet you need to influence to win more visibility on the queries you track. ## Left: Sources used to talk about *your brand* **What it is:** The citations AI uses in responses where your brand is discussed or referenced. Think of this as the “source ecosystem behind your brand’s narrative.” **How to use it:** Use this panel to understand whether AI is learning about you from: * Your own website (best case) * Third-party editorial coverage * Competitor pages * Or social/community sources ### What it means if your website is the top source (high Owned share) If the left panel is dominated by **Owned** citations (your domain), it usually means: * Your site contains clear, citeable information about your product and brand, * AI can reliably use your content as evidence, * your narrative is more controllable (you can update your pages and AI will pick it up). **Typical next step:** Keep strengthening a small set of “source pages” (pricing, security/trust, comparisons, FAQs) because these pages often drive both better rankings and more accurate sentiment. ### What it means if your website is *not* the top source If the left panel is dominated by **Off-page**, **Social**, or **Competitor** citations, it usually means: * AI is learning about your brand primarily from outside sources, * your owned pages may be missing key information or not structured to be cited, * narratives (including negatives) may be shaped by third parties. **Typical next steps:** * Upgrade your owned “source pages” (Website → Pages) so they’re citeable (clear headings, direct answers, FAQs + schema). * Earn more high-quality mentions on trusted editorial domains (Off-page → Outreach). * If Social dominates, engage where those narratives form (Off-page → Engage). If your brand is being described inaccurately or negatively, the left panel is usually the fastest way to see where that narrative is coming from. ## Right: Sources used to answer *your prompts* The citations AI uses across answers to the prompts you track (whether or not you are mentioned). Think of this as the “source corpus” AI is pulling from to answer the exact questions you want to win. This panel is often the most important for improving visibility because it shows the sources that *define* the answers in your category. ### How to use it Use the right panel to answer: * Which websites dominate the category conversation for our prompts? * Are the driver sources mostly editorial, social, or owned brand pages? * Which specific URLs are being cited repeatedly for our key prompt themes? ### What to do with this information (the practical takeaway) If you want to show up more often, you typically need to do one (or more) of these: * **Get included in the driver sources** (Off-page Outreach to the domains that keep appearing). * **Publish content aligned to those sources and formats** (Opportunities / Content) so your pages compete with what AI already trusts. * **Create or upgrade “source pages” on your site** that directly answer the prompt intent (Website → Pages), then build authority so AI starts citing them. ### Why your site may be absent on the right panel (and why that’s normal) It is very common for the right panel to be dominated by: * Major editorial sites * Wikipedia-style references * Community sources like Reddit * Or niche review sites. That is not “bad” by itself. It simply tells you what the model is currently using as evidence for your prompt set. Your goal is to **earn a place in that evidence set** over time. A simple heuristic: * The **left panel** tells you whether AI is sourcing *your brand story* from the right places. * The **right panel** tells you what sources *you need to compete with or influence* to win more prompts. ## Domains vs URLs Citations typically have two levels: * **Domains** tell you which sites matter most (strategy view). * **URLs** tell you which exact pages are being cited (execution view). If you are deciding where to invest off-page effort, start with Domains. If you are creating a specific “source page,” start with URLs. ## Citation types Meridian groups citations into: * **Owned** (your domain), * **Off-page** (third-party editorial), * **Competitor** (competitor domains), * **Social** (forums and social platforms). This breakdown helps you understand what the ecosystem is rewarding. ## How to use Citations to plan work A practical approach is: 1. Identify top driver domains and URLs. 2. Find where competitor domains are heavily cited. 3. Decide how you can realistically influence the ecosystem: * Earn inclusion on third-party pages (Off-page Outreach). * Participate in communities where social sources drive answers (Off-page Engage). * Improve your own “source pages” so they are citeable (Website → Pages fixes). If Owned citations are low, the fastest improvements often come from upgrading a small number of pages into “source pages” with clear headings, direct answers, and FAQ + schema. ## Example use cases * If Reddit is a major driver in your category, it is often worth engaging in a small number of high-impact threads consistently. * If a specific editorial domain drives many answers, outreach for inclusion on those pages can materially change rank stability. * If competitor domains dominate, it is a signal that your “official” pages are not being relied on as sources yet. # Competitors Source: https://docs.trymeridian.com/analytics/competitors Compare performance over time and identify where competitors are consistently outranking you. Competitors helps you understand performance in context. AI answers are often list-based, and you rarely “win” in isolation; you win relative to the brands that appear in the same responses. Competitors tab showing visibility trend chart and competitor table ## Why Competitors matters Competitor insights are most useful when you’re trying to answer questions like: * “Who is taking the top slots for our most important prompts?” * “Are we losing visibility overall, or are we losing to one specific competitor?” * “Which competitor spikes correlate with our declines?” * “Which topics should we invest in next (content, off-page, website fixes)?” Competitors is best used as a prioritization surface. Once you identify who is winning and where, you typically jump to: * **Analytics → Prompts** (to inspect the exact runs), * **Analytics → Citations** (to see which sources support the winners), * and **Off-page / Website / Content** to execute. ## What you see on the Competitors tab ### 1) Competitor Visibility chart The chart shows competitor visibility over time for the selected date range. It is designed to help you spot trends and spikes. #### How to interpret the chart * If one competitor steadily rises while you drop, they are likely capturing more “recommendation space” for your prompt set. * If a competitor spikes on specific days, it’s worth checking what changed (new citations, new content, platform changes). * If all lines move together, that often indicates platform behavior or sourcing changes rather than a single competitor action. ### 2) Competitor table The table summarizes each competitor’s performance and where they tend to appear. You’ll typically see: * **Competitor** (brand) * **Top prompt** (the prompt where they show up most) * **Top topic** (the topic where they show up most) * **Visibility** (how often they appear) * **Prominence** (their average rank; lower is better) ## Filters (focus your competitive comparison) Everything in Competitors respects your filters. Use the controls at the top of the page to focus the comparison. For example: * specific **topics** or **prompt groups**, * a specific **AI platform/model**, * a specific **date range** (and Prev. Period), * or other segments you’ve set up. Filtering is the fastest way to answer questions like “Which competitors beat us on ChatGPT?” or “Who is winning in this one topic?” ## Adding competitors Click **Add competitor** to add brands you want to track against. Add competitors modal with search and selected list ### Best practices for competitor selection * Add **3–8 true substitutes** (brands customers would realistically choose instead of you). * Avoid adding non-competitors “just to monitor them.” Too many competitors can make trends noisy. * If you manage multiple categories, include competitors relevant to each category. Your competitor list affects other parts of Meridian. For example, Off-page Outreach filters out competitor domains based on this list. If a competitor shows up in Outreach, adding them here will filter them out going forward. ## Removing competitors If you want to remove a competitor, use the delete action in the table (or the competitor row actions, depending on your UI). You’ll see a confirmation modal. Removing a competitor: * stops tracking them in competitor charts and tables, * and updates downstream filters that depend on competitor set (for example, Outreach filtering). ## How to turn competitor insights into actions Competitor wins usually come from one (or more) of these drivers: ### 1) Source advantage (they are cited more often) If competitors are supported by strong editorial or social sources, they will often outrank you even if your owned content is good. **What to do:** * Go to **Analytics → Citations** and see which domains/URLs are driving answers for your prompts. * Use **Off-page → Outreach** to pursue inclusion in the same high-impact editorial sources. * Use **Off-page → Engage** if social sources are a meaningful driver in your category. ### 2) Format advantage (their content matches the prompt better) If your prompts are “best X” or “X vs Y” and competitors have better comparison structure, they often win. **What to do:** * Use **Opportunities** to generate the right content format (listicle, comparison, guide). * Use **Content** to produce and publish a piece that matches the prompt intent (criteria, “best for,” trade-offs). ### 3) Positioning advantage (their narrative matches criteria) If AI answers emphasize criteria like fees, trust, reliability, or customer support, competitors win when they have clearer proof and better narrative coverage. **What to do:** * Use **Analytics → Sentiment** to see which dimensions are driving narrative. * Improve “source pages” on your site (Website → Pages) so your differentiators are easy to cite. Competitors tells you “who is winning and where.” The fastest path to action is to jump from a competitor’s top prompt into run-level evidence in Prompts. # Products Source: https://docs.trymeridian.com/analytics/products See which products are recommended in AI answers and improve product-level outcomes. Products connects AI visibility to outcomes. It shows which products are being mentioned in AI answers and which never appear, so you can improve product-level structure and coverage. Products grid showing products marked Mentioned vs Not mentioned Products is most relevant for brands that sell physical products that can show up in AI shopping-style results (product cards). If your business is not product/SKU-based, you can ignore this section. Right now, Products supports **Shopify merchants** by importing SKUs from Shopify. We’ll be adding additional product catalog integrations over time. ## Products and prompts (important context) Products only shows visibility for the prompts you’re tracking. Meridian is monitoring whether each product appears in AI answers **for your tracked prompt set**. That means product visibility is driven by two things: 1. Whether AI answers include your products, and 2. Whether you are tracking prompts where a product could realistically be recommended. ### Example If you sell basketball shoes but you aren’t tracking prompts such as “best basketball shoes” or “best shoes for indoor basketball,” you will almost certainly see products marked **Not mentioned**—because Meridian has no relevant answers to evaluate. If you want reliable product monitoring, make sure you track prompts that match how people discover and compare products in your category. ## Recommended setup To make Products useful quickly, build prompt coverage around your catalog. 1. Create a **Topic** for each product line or category (for example: “Basketball shoes,” “Trail running shoes,” “Compression shorts”). 2. Then either: * auto-generate **10 prompts per topic**, or * add prompts manually for your highest-priority products and categories. Once those prompts are tracked, Products becomes a clear signal of which items are being recommended (and which are missing) for the queries that matter. ## What you see in Products ### Product cards (grid view) Each product appears as a card with: * product image and title * price (if available) * tags/categories (if available) * a status label: * **Mentioned** (your product appears in AI answers for the tracked prompts) * **Not mentioned** (it does not appear in the tracked answers) You’ll also typically see an action like **See X responses** on mentioned items so you can inspect where and how the product is showing up. ### Catalog count At the top right, you may see the number of products in your catalog. This helps confirm Meridian is reading the correct product set for your company profile. ## How to use Products (practical workflow) ### Step 1 — Start with “Not mentioned” products Scan for products marked **Not mentioned**. These are usually the highest-leverage gaps, especially if they are core to your business. Before you assume there is a product problem, sanity check: * Are you tracking prompts that would recommend this product or category? * Is the product name or category obvious in the prompt set? If prompt coverage is missing, fix prompts first. ### Step 2 — Inspect products that are Mentioned For products that are **Mentioned**, click **See responses**. This is where you learn *why* the product appears and what pattern you should replicate for other products. Use “See responses” to understand: * which prompts trigger recommendations, * which competitors/products appear instead (when you are not #1), * and what criteria the answer uses (price, durability, trust, fit, etc.). Product detail page showing runs that mention the product ### Step 3 — Learn the winning pattern (and copy it) When you find runs where your product is recommended, look for: * prompt format (best-of list, comparison, “for X use case”) * what criteria the answer emphasizes * whether the response is driven by citations (and which sources) * whether the answer is product-specific (specs/fit) or brand-level (trust/quality) Then apply those patterns to: * other similar products, * your category pages, * and supporting content. ### Step 4 — Prioritize “money products” and “money prompts” You will get the fastest results when you focus on: * your top revenue products, * the prompts most likely to drive purchase intent, * and the categories where your products are competitive. Products is most powerful when paired with your prompt strategy. ## What can improve product mentions Product improvements are usually structural and comparative. AI systems recommend products when the criteria are clear, the product fits the use case, and the information is easy to summarize. Common improvements include: * Add clear specs and “who this is for” sections. * Add comparisons against common alternatives (and be explicit about trade-offs). * Add FAQs based on the questions AI is trying to answer. * Ensure category/list pages include your products and match prompt formats. If your brand is mentioned but your products are not, the fix is usually product page structure and recommendation criteria coverage (not general brand awareness). ### Practical examples of high-impact upgrades * If prompts are “best X for beginners,” add beginner-oriented criteria (ease of use, fit guidance, sizing, care). * If prompts are “best X for durability,” add durability proof (materials, testing, guarantees). * If prompts are “best value,” add clear price/value framing and what’s included. ## Measuring impact Over the next period, you should see: * more products moving from **Not mentioned → Mentioned**, * improved prominence on prompts where product recommendations appear, * and more citations that reference your owned product/category pages. If nothing moves after a full period, the most common reasons are: * you are not tracking the right product/category prompts, * competitor sources dominate citations for those prompts, * or your product/category pages are not structured in a citeable way. # Prompts Source: https://docs.trymeridian.com/analytics/prompts Track the questions you care about, organize them into Topics, and diagnose performance down to the individual AI responses. Prompts is the foundation of Meridian. Everything you measure in Analytics starts with the prompts you track. If you track the right prompts, Meridian can tell you where you are being recommended, where competitors are winning, and what to do next. This page covers the Prompts tab end-to-end, including how to add and organize prompts, and how to use **Analyze** to inspect platform breakdowns, run-level results, and the full response (including citations and shopping output when available). Prompts tab showing the prompts table with key metrics and controls ## What prompts are (and why they matter) Prompts are the questions Meridian tracks across AI platforms. They define what you are measuring. A helpful way to think about prompts is: **“What are the queries we want to show up more for in AI search?”** Meridian reports on (and help you improve visibility) for the prompts you track. If you track the right prompts, you get a clear view of: * where you are being recommended today, * where competitors are winning, * and which actions (content, off-page, website fixes) are most likely to increase visibility over time. If you track the wrong prompts, the dashboard can look “good” without actually measuring discovery. ## How often prompts run (daily) Meridian runs each tracked prompt **once per day** across the AI platforms/models enabled in your account. This daily cadence is important because AI answers can vary day to day, so you want enough repeated runs to see real trends instead of one-off noise. ### Choosing the right prompts matters (a lot) It is completely normal, and often more valuable, to add prompts where you currently have no or low visibility, especially higher-funnel “best X”, “top X”, and comparison prompts. Try to avoid building a prompt set that is mostly: * **too branded** (e.g., “Is \[Your Brand] good?”, “What is \[Your Brand]?”), or * **too company-specific** (queries only someone who already knows you would ask). Those prompts often produce artificially high visibility because the assistant already expects your brand to appear. That can make performance look strong while not actually measuring discovery. A healthy prompt set usually includes: * discovery prompts (“best \_*\_”, “top \_*\_”, “\_\_\_ alternatives”), * comparison prompts (“\_*\_ vs \_*\_”), * use-case prompts (“\_*\_ for beginners”, “\_*\_ for \[specific need]”), * and a smaller set of branded prompts for monitoring reputation. ## What you can do in Prompts Prompts is designed for four core workflows: 1. Build a prompt set that matches how customers ask AI for recommendations. 2. Organize prompts into Topics so you can evaluate performance by category/product line. 3. Tag prompts with Attributes so you can filter and segment. 4. Diagnose performance using Analyze and run-level evidence. ## How the Prompts table is structured Each row in the table is a prompt you are tracking. For each prompt, Meridian shows: * **Topic**: the category the prompt belongs to. * **Visibility score**: how often your brand is mentioned for this prompt. * **Share of voice**: your share of mentions compared to tracked brands. * **Prominence**: your average rank when mentioned (lower is better). These prompt-level metrics roll up into Home and into topic-level summaries when you group by Topic. ## Adding prompts (three ways) Click **Add prompts** to open the menu. You can add prompts in three ways: 1. **Add Prompt** (one at a time) 2. **Add Topic** (create a topic and optionally auto-generate prompts) 3. **Upload CSV** (bulk import) ### 1) Add Prompt Use Add Prompt when you are adding a handful of high-value prompts manually. You can set: * the prompt text, * the topic, * optional attributes. ### 2) Add Topic (and auto-generate prompts) Use Add Topic when you want to scale quickly. A topic is usually a category or product line. You can optionally enable **Generate 10 relevant prompts for this topic**, which is a fast way to create a baseline prompt set for that category. ### 3) Upload CSV Use Upload CSV when you already have a prompt library and want to import at scale. The upload modal lists required and optional columns. In the UI shown: * Required columns: `prompt`, `topic` * Optional columns: `attributeId`, `attributeValue` ## Time range and previous period comparison At the top of the Prompts tab you can set: * a **date range** (for example: Last 7 days), * and compare against **Prev. Period**. This helps you answer “are we improving week-over-week?” without manually tracking numbers. ## Filters (scoping your analysis) Filters narrow the dataset you are looking at. Use filters when you want to answer questions like: * “How are we doing on ChatGPT only?” * “Which prompts in Topic X are consistently missing us?” * “How are we doing on a specific segment of prompts (attributes)?” If the table looks empty after filtering, widen your time range first and then remove filters one-by-one. Empty states are most often caused by overly narrow filters or small sample sizes. ## Group by: Topic (and why it matters) Prompts supports grouping so you can evaluate performance at a higher level than a single prompt. Prompts Grouped **Group by: Topic** is the most common view because it answers: * “Which categories are we winning?” * “Where are we consistently missing or ranked low?” * “Which topic should we invest in next?” When grouped, each topic section shows a rollup of: * visibility score, * share of voice, * prominence, plus the prompts inside the topic. ### How to use topic rollups * If a topic has strong visibility but weak prominence, you are being included but not preferred. That usually calls for better comparisons, proof, and citations for that topic. * If a topic has low visibility, you likely have a coverage or authority gap for that category. ## View settings View settings lets you change how the table is displayed (for example, which columns you see). Use it to keep the table focused on the metrics your team cares about. ## Prompt usage and limits At the top right you may see a usage indicator like “97 / 300 prompts.” This shows how many prompts you are tracking relative to your plan limit. ## Exports Use **Export** when you want to share performance with teammates or archive a reporting snapshot. ## Attributes (segmentation) Attributes allow you to tag prompts so you can filter and compare different slices of your prompt set. Examples: * Intent: Research vs Decision * Audience: Beginner vs Expert * Internal: Campaign tags Click **Attributes** to manage and create attributes. Manage attributes modal showing system and custom attributes ### Creating a new attribute You can create custom attributes with types like: * Text * Number * Single select * Multi select # Analyze (run-level evidence) Analyze is where you stop looking at aggregate scores and inspect what actually happened in the underlying AI responses. Use Analyze when: * a prompt is important and you want to understand why you are winning or losing, * you see a sudden change in visibility or prominence, * you want to compare platform performance (ChatGPT vs Gemini vs Perplexity), * or you want to see which sources are being cited in the answers. Analyze view showing platform breakdown, brand visibility trend, and run table ## How to open Analyze From **Analytics → Prompts**, click **Analyze** on a specific prompt (or open the prompt row if your UI opens analysis directly). Once you open Analyze, you’ll see three layers of detail: 1. platform-level performance, 2. trend over time, 3. and the run table (the evidence). Analyze(single) ## Platform Breakdown The Platform Breakdown chart shows how performance differs by platform. This matters because different platforms rely on different sources and can return different ranked lists. ### How to interpret Platform Breakdown * If one platform is much lower, you often need a platform-specific plan (usually citations/off-page). * If one platform is high and others are low, your content and source ecosystem likely match the “high” platform better than the others. A practical next step is to open runs from the weaker platform and inspect which sources are cited. ## Brand Visibility trend The Brand Visibility chart shows visibility over time for the selected prompt. It helps you spot: * when a drop began, * whether changes are consistent or noisy, * and whether there are clear competitor-driven shifts. ### How to use the trend * A one-day dip is often noise. * A sustained dip across multiple days is usually a real change worth investigating. ## Response table (the evidence) The response table shows each run for the prompt, including: * **Date** * **Mentioned** (Yes/No) * **Position** (if mentioned) * **All mentions** (icons showing which brands/products appeared) * **Response** snippet * **Platform** * **Region** ### How to interpret the run table * If you have many **Not mentioned** rows, you have a coverage/authority gap for this prompt. * If you are mentioned but positions are **#4+**, you are included but not preferred. That often points to positioning, proof, or citation gaps. * If performance varies widely by platform, focus your investigation on the weaker platform’s runs first. ## Opening a response (Mention details) Click a response row to open the detailed response modal. This is where you can see the full answer, mentions, shopping output (if available), and citations. Mention details modal with mentions at the top and tabs for Response/Shopping/Citations ### What you see at the top * Prompt name * Properties (platform, date, region) * Mentions list (ranked entities) This tells you immediately: * whether you were included, * who outranked you, * and which competitors appear in the same answers. ## Response tab The Response tab shows the full answer text. This is useful for understanding: * what criteria the answer uses, * how the answer frames the category, * and what “proof” it relies on (often reflected in citations). ## Shopping tab (when available) Some prompts produce a shopping-style output, such as products with images, prices, and retailers. The Shopping tab shows that view when it is present. If Shopping output exists for your category, it can change the strategy: * product page structure matters more, * and citations/sources can influence which products appear. ## Citations tab The Citations tab shows the sources referenced in the answer (domains/URLs). This is often the most actionable part of a run because it tells you what the model trusted when forming the answer. ### How to use citations in a losing run If a competitor outranks you and the answer cites a specific editorial page or domain, that source is often a high-leverage target for: * **Off-page → Outreach** (earn inclusion), or * improving your own source pages so you can be cited instead (**Website** **→** **Pages**), or * publishing content aligned to what the source covers (**Opportunities** / **Content**). Run-level citations often explain why prominence is low even when visibility is high. If your pages are not in the citation ecosystem, it is difficult to hold top ranks consistently. # Sentiment Source: https://docs.trymeridian.com/analytics/sentiment Understand how AI describes your brand and fix narratives that suppress rankings. Sentiment helps you understand whether AI is describing your brand in a way that helps or hurts recommendations. Even when you are mentioned, recurring objections (fees, reliability, trust, customer service) can push you down the list and reduce prominence. The goal of Sentiment is not to “sound better.” The goal is to identify the specific narratives that show up repeatedly in AI answers for your prompts, then address them with clarity and evidence on pages AI can cite (and in off-page sources when needed). Sentiment ## When to use Sentiment Sentiment is most useful when: * Your **prominence worsens** even though visibility is stable (you’re included, but not preferred). * You notice a drop in the **Sentiment** KPI on Home. * You want to understand why a competitor is consistently ranked above you (the answer criteria often includes narrative signals). * You suspect a specific objection is hurting you (pricing, reliability, trust, customer support). ## How to read the Sentiment page ### Overall sentiment score At the top you’ll see an overall **Sentiment Score** for the selected time range, with a period-over-period change. This gives you a directional read: “Is AI describing us more positively or more negatively than last period?” ### Dimension tiles (the actionable part) Below the overall score, you’ll see dimension scores (for example: Security, Reputation, Ease of use, Features, Reliability, Customer service, Fees, Price, Quality, Performance). These are usually more actionable than the overall score because they tell you what AI repeatedly emphasizes. If one dimension is much lower than others (for example, Fees or Customer service), it is often the best place to start. ### Sentiment Breakdown chart The Sentiment Breakdown visualizes positive/neutral/negative mentions by dimension. * Green indicates positive mentions. * Gray indicates neutral mentions. * Red indicates negative mentions. This helps you answer: “Which dimension is driving negative narrative most often?” ## Filters (focus your analysis) Everything in Sentiment respects your filters. Use the controls at the top of the page to narrow what you’re inspecting—for example: * specific **topics** or **prompts**, * a specific **AI platform/model**, * a specific **date range** (and Prev. Period), * or other segments you’ve set up. Filtering is the fastest way to answer questions like “Is this a ChatGPT-only issue?” or “Is sentiment negative only for one topic?” ## Drilling into a dimension Click a dimension (for example, **Security**) to see the run-level evidence behind it. This drilldown is where Sentiment becomes actionable, because you can review: * which runs contained negative mentions, * which platforms those mentions came from, * and what the responses actually said. Dimension drilldown table for a specific dimension # Brand Kit Source: https://docs.trymeridian.com/brand-kit Define brand facts and voice so generated briefs and articles stay accurate, specific, and consistent. *** Brand Kit is where you give Meridian the context it needs to write about your brand correctly. When Brand Kit is filled out, briefs and articles are typically less generic, your positioning is clearer, and the content is less likely to include inaccurate or overconfident claims. Brand Kit is especially important if you use **Content** to generate briefs/articles, because the system will use this information to decide what to emphasize (and how to say it). Brandkit If Brand Kit is vague, generated content will be vague. If Brand Kit is specific and evidence-backed, generated content will be specific and evidence-backed. ## What Brand Kit affects Brand Kit is primarily used in: * **Brief generation** (what angle to take, what to emphasize, what to avoid) * **Article generation** (tone, terminology, claims, and examples) * **Refinements** (keeping updates consistent when you request changes) It can also help with: * consistency across writers and stakeholders, * reducing risky phrasing in regulated or sensitive categories, * and improving how clearly your differentiators show up in “best X” and comparison-style content. ## How Brand Kit is organized Brand Kit typically includes two kinds of inputs: 1. **Structured fields** (quick, high-signal inputs like mission and voice) 2. **Knowledge Base** (longer context you want Meridian to reference like competitive positioning, differentiators, policies, and proof) ## 1) Fill out the core fields (fast, high leverage) ### Brand mission Write 1–2 sentences describing what you do and why you exist. Keep it specific. Good: * “We help teams measure and improve how AI assistants mention, rank, and cite their brand and products.” Too vague: * “We help businesses grow.” ### Ideal customer profile Describe who you are for in plain language (role + company type + use case). Examples: * “In-house growth and SEO teams at DTC brands.” * “Marketing teams at SaaS companies that care about AI discovery.” ### Core values Add 3–6 values that should show up in tone and framing (for example: accuracy, transparency, customer-first, safety). These fields are short on purpose. They should capture the core identity you want repeated consistently across briefs and articles. These fields are short on purpose. They should capture the core identity you want repeated consistently across briefs and articles. ## 2) Set your Voice & Style The “Voice & Style” section helps Meridian match your writing style so generated content sounds like you. ### Author persona Describe how the writing should feel. Examples: * “Friendly expert. Short sentences. Practical. Minimal hype.” * “Direct and technical. Focus on definitions, steps, and evidence.” ### Tone of Voice Concrete rules are more useful than adjectives. Good rules: * “Lead with a direct answer in the first 2–3 sentences.” * “Use bullets for criteria lists.” * “Use comparison tables when writing comparisons.” * “Avoid vague superlatives unless you include proof.” ## 3) Add Writing Samples (strongly recommended) Writing samples teach Meridian the “shape” of your best writing: tone, structure, headings, and how you explain things. ### What to include Add 2–5 samples: * one blog post (your standard tone) * one product/feature explainer (how you describe your offering) * one comparison-style piece (if you have it) * one trust/policy page (if your category is regulated) ### Tips * Use examples that you would be happy publishing again today. * If a sample is outdated or off-brand, do not include it. * Shorter and clearer samples tend to work better than long, meandering ones. ## 4) Knowledge Base (where you add deeper context) The Knowledge Base is the most flexible part of Brand Kit. This is where you can add detailed information that doesn’t have a dedicated UI field. Examples of what belongs in Knowledge Base: * Competitive positioning (who you compete with and how you differ) * Key differentiators (with proof, links, and “best for” fit) * Approved claims / prohibited claims (compliance boundaries) * Pricing and packaging details (what’s true, what’s nuanced) * Product specs, feature lists, integration details * Customer proof (case studies, testimonials, benchmarks) * Brand guidelines PDFs ### Text vs PDF When you add knowledge, you can choose: * **Text**: best for structured bullet lists, rules, and frameworks you want Meridian to quote cleanly * **PDF**: best for existing brand guidelines, one-pagers, policies, or longer documents you already maintain If you want Meridian to consistently reflect your differentiators or competitive positioning, put them in the Knowledge Base as a Text item. It is easier for the system to reuse clean bullet rules than to infer them from marketing pages. ### Recommended Knowledge Base items (templates) #### A) Competitive positioning (Text) Use this format: * **Competitors we’re compared to:** A, B, C * **When we win:** (3–6 bullets with criteria) * **When we’re not the best fit:** (2–4 bullets; honest trade-offs help) * **Language to use:** (preferred phrasing) * **Language to avoid:** (phrasing that is misleading or risky) #### B) Differentiators + proof (Text) * Differentiator * Why it matters * Proof (URL or document reference) Example: * “SOC 2 audited” → “Trust / compliance” → link to trust page #### C) Claims and boundaries (Text) * **Approved claims:** (bullet list) * **Disallowed claims:** (bullet list) * **Claims that require exact wording:** (bullet list + required phrasing) This is especially useful in regulated categories. #### D) Product/feature facts (Text or PDF) * A canonical feature list * Integration support * Constraints/limitations (so content doesn’t overpromise) ## How to keep Brand Kit useful Brand Kit is most effective when it stays current. Recommended maintenance: * Update it when pricing, product features, or policies change. * Add new proof points when you publish authoritative pages. * Add a Knowledge Base note when you notice repeated inaccuracies in generated drafts. * Remove or replace writing samples that no longer match your voice. If you have multiple Knowledge Base items that contradict each other, Meridian may produce inconsistent output. Keep one “source of truth” doc per topic (positioning, pricing, claims). ## Common mistakes (and how to avoid them) * **Being too generic:** “We are the best platform.” Replace with concrete criteria and proof in Knowledge Base. * **No boundaries:** If you don’t define prohibited claims, drafts may use overly confident marketing language. Add a “Claims and boundaries” Knowledge Base item. * **No trade-offs:** AI answers often rank by trade-offs. Include “when we’re not the best fit” in your competitive positioning. * **Old writing samples:** If samples are outdated, drafts will inherit outdated tone and structure. Refresh samples as needed. # Content Source: https://docs.trymeridian.com/content Create, review, and manage briefs and articles you can publish to improve AI visibility over time. Content is where Meridian turns “what we should do next” into publishable work. It is designed to help you create pages that are more likely to be **mentioned**, **cited**, and **ranked highly** in AI answers. Briefs and drafts are built around your tracked prompts, the sources AI already relies on, and best-practices for writing AI-cited content. Content list showing content pieces, briefs, and current statuses ## How Content is organized Content is built around two objects: * A **Brief** is the plan. It captures the target intent, recommended structure, and the evidence/citations that matter for the prompts you want to win. * An **Article** is the draft output generated from the brief. You can edit it, refine it, and publish it to your CMS. Most teams follow a simple flow: 1. Create a brief (from Opportunities or manually) 2. Review the brief 3. Generate an article 4. Refine and publish 5. Measure impact in Analytics ## The Content list (columns and statuses) The main table is your content pipeline. Each row is one piece of content in progress. ### Columns * **Type**: The format (Blog Post, Listicle, How-to Tutorial, Ultimate Guide, Comparison Post). * **Opportunity**: If the item came from Opportunities, this shows the originating recommendation. If you created it manually, this may be blank or reflect the manual title. * **Brief**: A shortcut to open the brief (typically View brief). * **Content**: The next step or action (Generate article, Review content, Researching, Add context). * **Created at**: when the brief was created. ### What the Content status means * **Generate article** The brief is ready. Clicking this will create a full draft article from the brief. * **Review content** The article is drafted and ready for review. * **Researching** Meridian is still assembling inputs for the brief/article. * **Add context** Meridian needs additional inputs before it can generate a high-quality output. This typically happens when your prompt/topic is broad or when the system cannot confidently infer the right angle, constraints, or sources. Adding a bit of context here usually improves quality significantly. The “X / Y articles” indicator in the top right reflects your plan usage (how many article generations you have used out of your plan limit). ## Creating a new brief You can create briefs in two ways: 1. From **Opportunities** (recommended when you want the highest-impact suggestions), or 2. Directly from **Content** using **New brief** (recommended when you already know what you want to write). ### Option A: Create a brief from Opportunities Opportunities is a good starting point when you want Meridian to pick topics that are most likely to increase visibility and citations for your tracked prompts. A typical flow: * Open **Opportunities** * Pick a High (or Medium) item * Click **Generate brief** * Return to **Content** to find the new row and click **View brief** ### Option B: Create a brief directly in Content Click **New brief** to open the content type chooser. Create new content modal with content format options You’ll see common formats: * **Smart suggestion (Suggested)**: Meridian chooses the best content type based on what sources are most cited for your selected topics/prompts. * **General blog post**: Best for informational topics where you want a clear intro + sections + conclusion. * **Listicle**: Best for “best X” prompts and prompts that return a list of recommendations. * **How-to tutorial**: Best for step-by-step prompts (“how do I…”). * **Ultimate guide**: Best for broad topics that need comprehensive coverage. * **Comparison post**: Best for “X vs Y” prompts or “best alternative to…” prompts. ### Smart suggestion: choosing Topics vs Specific Prompts If you select **Smart suggestion**, Meridian will choose the format we believe is most-likely to earn citations for the given prompt or topics. Smart suggestion selector: Topics vs Specific Prompts * **Select by Topics** Choose one or more Topics. Meridian will look at the prompts inside those topics and generate briefs that best match the citation ecosystem and content gaps. * **Select Specific Prompts** Choose the exact prompts you want to write for. This is useful if you have a short list of “money prompts” you are actively trying to win. Once you select topics/prompts, click **Create briefs**. Meridian will then run the brief generation process. ## Reviewing a brief (what to check before generating an article) Opening **View brief** shows the brief editor. This is where you validate that the plan matches your goal before generating a full draft. Brief review page with analysis, target prompts, and citations ### Target Prompts (right rail) The **Target Prompts** panel shows which prompts this brief is designed to win and your current visibility on those prompts. This is important because content only improves visibility if it maps to the prompts you are actually tracking. Practical checks: * Are the target prompts the ones you care about? * Are they in the right topic/category? * Are you missing prompts that should obviously be included for this content? ### Citations (right rail) The **Citations** panel shows the sources that are currently winning visibility for this prompt set. This is useful because: * It tells you which sources AI already trusts for this query * It helps you align your content to that ecosystem (structure, evidence, comparable claims). If the citations panel is dominated by a few domains, that is a strong signal those domains are shaping the category answer space. ### Recommendations (right rail) Recommendations are small, concrete improvements Meridian suggests for AI visibility. For example: * “Turn title into a question” * “Atomic intro answer” * other structural/citeability improvements Brief quality matters. A 2-minute review before generating an article usually saves you a lot of editing later and produces content that performs better in AI answers. When you are happy with the brief, click **Generate article**. ## Reviewing and refining the article After generation, you’ll land in the article editor. Article review editor with AI refinement input and History ### Editing the draft (what usually matters most) Before publishing, review the draft for: * Correctness (facts, claims, pricing, availability) * Specificity (examples, numbers, concrete criteria) * Structure (headings that match the prompt intent) ### AI Refinement (targeted edits) At the top of the editor you can request changes via the refinement input (e.g., “Rewrite in a more casual tone and add 3 competitor alternatives”). This is most useful when you give a specific instruction: * “Add a comparison table with these criteria…” * “Rewrite the intro to answer the question in 2 sentences…” * “Add an FAQ section with 8 questions about…” * “Add evidence/examples for claims in section 3…” ### Score / SEO / History On the right rail you’ll see tabs such as: * **Score** (quality/structure signals) * **SEO** (SEO-related checks) * **History** (audit trail of edits and refinements) ### Copy to publish When you are ready, click **Copy** and paste into your CMS. Treat generated content as a draft. You should always review for factual accuracy, compliance constraints, and brand voice before publishing. ## Measuring impact after publishing Content work should be measured against the prompts it was designed to influence. A simple measurement loop: 1. In **Analytics → Prompts**, watch visibility and prominence for the target prompts. 2. In **Analytics → Citations**, watch whether your owned pages begin to appear more frequently as citations. 3. On **Home**, watch whether KPI changes persist over the next period. If you published content but nothing moves after a full period, the bottleneck is often one of: * You are not in the citation ecosystem yet (Off-page work needed) * Your pages are not citeable (Website → Pages fixes needed) * Or you are not tracking the prompts that would surface that content (prompt set mismatch) Changes are not immediate, and not all content will lead to an increase in visibility.\ \ Quality of content is significantly more important than quantity of content. # Directories Source: https://docs.trymeridian.com/guides/dashboard/off-page/directories Get listed on the directories AI already trusts in your industry. Directories is a prioritized list of directory pages where AI assistants frequently pull evidence for your tracked prompts, but where your brand is not currently included (or not being cited). These are often high-leverage wins because many “best X” and comparison answers rely on directories as structured, trusted sources. Directories is based on what AI is already citing for your prompt set. It updates weekly so you’re always working from the current citation ecosystem. Some directories are broadly relevant across industries (for example, major marketplaces or review platforms). Others are highly industry-specific (for example, legal directories like superlawyers.com). Meridian surfaces the directories that matter most for *your* prompt set and citation ecosystem. The goal is not just to “get a link.” The goal is to add accurate, differentiated, on-brand language about your company in places AI already trusts—so that language shows up in AI answers over time. ## How Directories works Directories is built from the same citation ecosystem as Outreach and Engage: * Meridian starts with the sources AI is citing for your prompts (**Analytics → Citations**). * It identifies the **directory sources** that are driving answers in your category. * It filters down to the directories where: * you are not currently included, or * your presence is incomplete / not strong enough to be cited. Like the other Off-page tabs, this list is meant to be an execution queue you can work through over time. ## Refresh cadence (weekly) Directories refreshes weekly to reflect the past week’s citation data. As new directories become important in your category (or as your prompt set changes), your targets will shift. ## How to use Directories ### Step 1. Start with the highest-impact directories Not all directories matter equally. Prioritize the ones that show up repeatedly in the citation ecosystem for your tracked prompts. A good weekly cadence is to pick 3–10 directory targets, depending on how much time you have. ### Step 2. Confirm eligibility and how inclusion works Open the directory and confirm: * the directory is relevant to your category and geography, * the directory accepts listings (or has a submission flow), * and the listing page format matches how AI might cite it (clear fields, structured descriptions). Some directories are easy to join. Others require verification, paid tiers, or editorial approval. ### Step 3. Add language that matches how AI answers When you create or update a listing, focus on the language AI uses in recommendation answers: Include: * who you are and who you’re for (ICP), * your core offerings (products/services), * your differentiators (with specifics), * proof signals (certifications, compliance, ratings, awards, customer count, etc.), * and clear “best for” positioning (who should choose you). Avoid: * vague superlatives (“best,” “#1”) without proof, * marketing fluff, * and long paragraphs that don’t say anything concrete. A directory listing is often treated like structured evidence. Short, specific, criteria-based language tends to work better than long brand narratives. ### Step 4. Keep the listing consistent with Brand Kit Before you submit, check: * your exact positioning sentence, * approved vs disallowed claims, * and proof links. If you want Meridian-generated content (briefs/articles) to match your directory language, make sure those same differentiators and constraints exist in **Brand Kit** (especially Knowledge Base). ### Step 5. Track whether it starts showing up in citations After you are listed, measure impact week over week: * **Analytics → Citations**: does the directory domain begin appearing more often? Does your presence improve your owned citation ecosystem? * **Analytics → Prompts**: do you see improvements in visibility/prominence on prompts where directories influence answers? * **Off-page → Directories**: does the target fall off the list (because you’re now included)? ## Industry-agnostic vs industry-specific directories You will typically see two kinds of targets: ### Industry-agnostic directories These are broad platforms that cover many categories. They can be influential because they have strong authority and structured listings. Examples (varies by industry and prompt set): * major review platforms * marketplaces * general business directories ### Industry-specific directories These are often the highest-leverage directories because they map directly to how buyers research in your category. Example: * [**superlawyers.com**](http://superlawyers.com) for legal services If an industry-specific directory appears repeatedly in your citations ecosystem, prioritize it early. These are often “source of truth” sites for AI answers in that vertical. ## Important expectations Getting listed does not guarantee visibility increases. Many factors affect outcomes: * whether the directory is actually cited for your tracked prompts, * whether your listing contains the criteria the answer emphasizes, * whether competitors still dominate other core sources, * and whether your prompt set matches the queries that directory influences. The goal is to compound improvements over time by earning presence in the evidence sources AI already trusts. If you add your name but don’t add meaningful, differentiated language, you often won’t see impact. The content of the listing matters. # Social Source: https://docs.trymeridian.com/guides/dashboard/off-page/engage Social with the community pages AI is citing for your prompts to shape trusted language about your brand. Engage is a list of community and social pages (Reddit, LinkedIn, forums, etc.) that AI is actively using as sources for the prompts you track. Hero Dark The goal is not “social for social’s sake.” The goal is to add accurate, helpful language about your brand, products/services, and differentiators in the places AI already trusts so those narratives show up in AI answers over time. ## Why Social matters (and when it doesn’t) UGC and social sourcing is extremely industry-dependent. * In some categories (often DTC and consumer purchase prompts), AI relies heavily on community threads, reviews, and social discussion. * In other categories, AI relies mostly on editorial sources, official documentation, and company pages, and social sources rarely appear. The fastest way to tell which bucket you are in is **Analytics → Citations**: * If **Social** is a meaningful driver, Social is high leverage. * If Social rarely appears, Social may be secondary to Outreach and owned source pages. ## How the Social list is generated Social is built from the same citation ecosystem as Outreach, but filtered to social/community sources: * It starts with the social/UGC pages that are most cited for your prompt set. * It focuses on the pages that are already shaping answers (not generic “influencer lists”). Social is refreshed **weekly** to reflect the past week’s citation data. ## Two common ways teams use Social ### 1) Engage with existing posts that AI already trusts This is the highest-leverage use of Social. When you contribute to a thread that is already being cited for a prompt you care about, you are adding language directly into the evidence pool AI uses. A strong engagement contribution usually includes: * a short decision framework, * trade-offs and safety considerations, * concrete facts and specifics (not marketing language), * and clarifying context that matches what the user is asking. Linking to your website is often less important than the language itself. In many cases, the text you add (clear criteria, differentiators, factual claims) matters more than the link. ### 2) Use Social as inspiration for what to post next (and where) Social is also a research surface. The goal here is to learn what AI is already trusting from social/community sources, then use those patterns to publish better social content yourself (and improve your owned content where it matters). Use Social to learn: * what questions repeat, * what objections come up, * what language people use to compare options, * and what criteria the prompt seems to care about (price, durability, trust, ease of use, etc.). #### A) Turn this into your own social content plan Use those patterns to create a steady stream of posts on the platforms AI is already citing. In practice, teams use Social to decide: * **What to post**: the top recurring questions and comparison criteria (turn each into a post). * **How to frame it**: the exact language and trade-offs people use (mirror it, but add clarity and proof). * **Where to post**: the communities and platforms that show up in Social (those are the ones most likely to influence answers for your prompt set). Example content formats that work well: * “Decision framework” posts (“If you care most about X, choose Y”) * “Trade-off” posts (“Cheap vs secure vs fast—what you give up”) * “Myth vs reality” posts (address the most common objections directly) * Short comparison posts (you vs competitor, with clear criteria) The goal is not to drop links everywhere. It’s to get accurate, differentiated language into the places AI already trusts. #### B) Use the same patterns to improve your owned content too The same themes you see in Social are often exactly what your website needs to cover to become more citeable. Use those patterns to improve: * your FAQs and schema on owned pages (\*\*Website → Pages\*\*), * your content briefs and listicles (\*\*Opportunities / Content\*\*), * and your comparison criteria and “best for” positioning (so you rank higher when you are mentioned). In short: Social helps you learn what’s shaping answers today, then you can use it to publish smarter social content *and* strengthen the pages you want AI to cite. ## How to engage effectively 1. **Pick a page that is cited for a prompt you care about.** If the page is not in the citation ecosystem, it is usually less impactful. 2. **Be helpful first.** Focus on giving the best answer in the thread. Avoid promotional tone. 3. **Use specific language that matches AI prompt criteria.** If the prompt is “best X for beginners,” your comment should address beginner criteria (ease of use, learning curve, support, pricing clarity). 4. **Reference a link only when it genuinely helps.** A link is optional. If you include one, link to a source page that supports your claim. ### Example engagement approach If the thread is “best exchanges for beginners,” a strong comment includes: * a short decision framework (what to optimize for), * trade-offs (fees vs simplicity vs security), * and factual differentiators that are easy to repeat in an AI answer. ## Setting expectations Social is a leverage play, not a guarantee. Adding your brand name to a thread does not automatically increase visibility. Outcomes depend on factors like: * whether the page continues to be cited for your prompt set, * whether your contribution aligns with the criteria AI uses in answers, * whether competitors still dominate the cited sources, * and whether your owned pages are citeable when AI wants proof. Think of Social as “shaping trusted language.” The best results come from consistent, high-quality participation on a small set of high-impact pages over time. ## Refresh cadence (weekly) Social targets update weekly based on the past week’s citation data. As new threads trend and old ones fade, your target list will change. ## Measuring impact You will usually see impact in this order: 1. More inclusion or visibility in the engaged threads/pages. 2. More appearances in citations (Analytics → Citations) if those pages remain drivers. 3. Better prominence on prompts those sources influence. 4. Visibility increases across related prompt clusters. # Editorial Source: https://docs.trymeridian.com/guides/dashboard/off-page/outreach Identify the editorial pages AI is citing for your prompts and request inclusion where it matters most. Editorial is a list of third-party editorial pages that AI is already using as sources for the prompts you track. The goal is to earn mentions in the pages that shape answers in your category so that AI is more likely to recommend you over time. Outreach Editorial is not a generic PR list. It is built from citation data, which means the targets are biased toward sources AI already trusts for the exact queries you care about. ## What Editorial is (and why it matters) For many “best X” and comparison prompts, AI answers rely heavily on editorial lists and review content. If you are missing from those pages, you can be: * Mentioned less often. * Cited less often (or not at all). * And ranked lower, even if your website content is strong. Editorial turns this into a concrete execution list so PR work becomes trackable and prioritized. ## How the Editorial list is generated Meridian starts with the citation ecosystem in **Analytics → Citations** and filters it down to the most actionable editorial targets. ### 1) Start from editorial content that is being cited Editorial begins with editorial pages and domains that are actively being cited across your tracked prompt set. These are sources AI is already relying on to answer your prompts. ### 2) Filter out editorial where you are already included If you are already mentioned on a page/domain, Meridian filters it out so you do not waste time requesting inclusion where you already have presence. ### 3) Filter out competitor domains Competitor domains are not realistic editorial targets, so they are filtered out. If you see a competitor showing up in Editorial, it usually means Meridian does not recognize them as a competitor in your workspace yet. Add the competitor in **Analytics → Competitors** and they will automatically be filtered out going forward. ## Refresh cadence (weekly) Editorial is updated **weekly** to reflect the **past week’s citation data**. As AI citation patterns shift (new pages get cited, publishers update lists, new content is published) your targets will shift too. If you earned inclusion on a target editorial, you should often see it disappear from Editorial over time (because it becomes “already included”), and you’ll see the next best targets rise. ## How to use Editorial A practical workflow is: 1. **Start with targets that have high citation impact.** Prioritize domains/pages that appear repeatedly in the citation ecosystem for your prompts. The list is already priority-ranked by citation count (how often the piece of content was trusted & cited by AI over the previous week). 2. **Open the article and confirm the gap is real.** Confirm that a competitor is present and you are missing, and confirm the page is still up-to-date and relevant. 3. **Find the right submission path.** This can be an author email, editorial contact form, “write for us” guidelines, or a partner program. 4. **Pitch inclusion with clear evidence.** Your pitch should include: * Who you are. * Why you belong in the list. * What differentiator matters. * And a credible source page on your site that supports the claim. ### Example outreach pitch (short) “Hi \[Name] — I noticed your article covers \[category] and includes \[competitor]. We help \[audience] with \[use case], and we’re a strong fit for \[criteria]. Here is a concise source page with details and proof: \[URL]. Would you consider adding us as an option for \[audience/use case]?” ## Editorial depends on content type (effort varies) Not all targets are equal. The right outreach motion depends on what kind of page you’re trying to influence: * **Easy inclusion requests** Some pages are designed to accept updates (e.g., “best tools” lists with a contact form or clear editorial workflow). These are often quick wins. * **Harder, long-term targets** Some publishers are harder to influence and require relationship-building. In those cases, the realistic goal is often: * Being included in a future update of the article. * Contributing a quote or data point. * Co-authoring a piece on that domain. A good rule of thumb is to treat Editorial like a pipeline. Mix quick wins with longer-term targets so you steadily improve your citation footprint. ## Status tracking Use status to keep the list actionable: * **Todo** → you haven’t contacted yet * **In progress** → you reached out and are following up * **Done** → you are included or the target is closed Off-page work compounds. A small set of consistent editorial targets each week often improves citation composition more than one large burst. ## Measuring impact You will usually see impact in this order: 1. More inclusion on target editorial pages/domains. 2. More appearances in citations (Analytics → Citations). 3. Better prominence on the prompts those sources influence. 4. Visibility increases across related prompt clusters. Adding your name does not guarantee visibility increases. Outcomes depend on whether the page continues to be cited for your prompts, whether your mention aligns with the criteria the answer uses, and whether competing sources still dominate the answer. # Contact Support Source: https://docs.trymeridian.com/guides/settings-and-support/contact-support Get help fast by reaching out through email or directly in the Meridian dashboard. There are two primary ways to contact Meridian support: ## 1) Email You can always email us at: [**support@trymeridian.com**](mailto:support@trymeridian.com) Email is best when you want to: * include longer context, * attach files (exports, screenshots, docs), * or loop in teammates on the thread. ## 2) In-app support form (recommended) You can also contact support directly from the dashboard using the in-app **Contact Support** form. This is best when you want to: * report an issue while you’re looking at it, * quickly share the page you’re on, * and keep everything tied to your workspace. Contact Support form in the Meridian dashboard ### What to include (so we can help quickly) Whether you email or use the form, including the following details usually speeds things up: * **What you were trying to do** (example: “Export citations for last 7 days”) * **Where you were in the product** (example: “Analytics → Citations”) * **Date range + any filters** (platform, region, topic, attributes) * **Example prompt(s)** (if relevant) * **Screenshots** (especially of the error or unexpected output) * **Any exports** you generated (if relevant) If the issue is about a specific AI response, include a run where you can see: * whether you were mentioned, * your position, * and which sources were cited. # Glossary Source: https://docs.trymeridian.com/guides/settings-and-support/glossary Definitions for Meridian metrics, terms, and product concepts (kept consistent across the docs). This glossary is the single source of truth for Meridian definitions. If you see a term in the dashboard or docs, it should match the definition here. Most metrics in Meridian are scoped to your tracked prompts and the filters you apply (date range, platform, region, topics, attributes, and company profile). ## Core objects ### Company Profile A Company Profile represents a brand or client inside your organization. Each profile can have its own integrations (Google Analytics, Vercel), its own search terms, and its own datasets. You manage profiles in **Settings → Company Profiles**. ### Organization An Organization is the workspace container that holds company profiles, members, billing, and settings. ### Prompt A prompt is a tracked question you want to win in AI assistants. Prompts should represent real user discovery and evaluation behavior (for example: “best \_*\_ for \_*\_” or “\_*\_ vs \_*\_”). Prompts live in **Analytics → Prompts**. ### Topic A topic is a grouping mechanism for prompts (usually a product line, category, or theme). Topics help you organize and scale tracking. ### Attribute An attribute is a tag/segment you can apply to prompts to filter and group analysis (for example: “Beginner,” “Enterprise,” or “High intent”). Attributes are used primarily in **Analytics → Prompts**. ### Run A run is one execution of a prompt on a specific platform and region in a specific time window. Runs are the unit Meridian uses for evidence. Runs include: * whether you were mentioned, * your position (if mentioned), * citations (if present), * and a response snippet/full response. ### Mention A mention means your brand (or product) appears in the AI response for a run. ### Position Position is your rank in a specific run when the answer is list-like (for example: #1, #2, #3). ### Prominence Prominence is your average position when you are mentioned. Lower is better. **Formula:** Prominence = average(position | mentioned) Prominence is shown on **Home** and also per prompt in **Analytics → Prompts**. ### Citation A citation is a source reference in an AI answer (a domain or URL). Citations are analyzed in **Analytics → Citations**. ### Domain A domain is the site-level source ([example.com](http://example.com)). Domains are typically used for strategy (“which sites matter most?”). ### URL A URL is the page-level source (example.com/page). URLs are typically used for execution (“which exact pages are being cited?”). ### Citation types: Owned / Off-page / Social / Competitor Meridian classifies citations into: * **Owned**: your domain(s) * **Off-page**: third-party editorial sites * **Social**: community/social platforms (Reddit, LinkedIn, forums) * **Competitor**: competitor domains ## Metrics (what they mean + how they’re calculated) ### Visibility (%) Visibility measures how often your brand is mentioned across runs in the selected timeframe. **Formula:** Visibility = (runs where you are mentioned / total runs) × 100 **Example:** 200 runs, mentioned in 96 → Visibility = 96/200 = 48% **Where it appears:** Home, Analytics → Prompts **How to interpret:** * Low visibility means you are frequently excluded from answers for your tracked prompts. * Visibility usually improves when you increase coverage and/or become present in the sources AI trusts. ### Share of voice (%) Share of voice (SOV) estimates how much of the recommendation space you occupy relative to competitors. **Formula:** SOV = (mentions of you / mentions of all tracked brands) × 100 **Example:** You are mentioned 90 times; all tracked brands total 500 mentions → SOV = 18% **Where it appears:** Analytics → Prompts Answer formats vary (some answers list 3 brands, others list 10). Use share of voice directionally and use a specific run’s details as the source of truth for exact ranked lists. ### Citation Rate (%) Citation Rate measures how often your owned domain is cited in answers. Specifically, it is the share of runs where the response contains at least one citation to your owned domain(s). **Formula:** Citation Rate = (runs with ≥1 citation to your domain / total runs) × 100 **Example:** 200 runs, your domain cited in 35 → Citation Rate = 17.5% **Where it appears:** Home KPI; details in Analytics → Citations **How to interpret:** * Higher citation rate usually means AI is relying on your pages as evidence more often. * Drops often mean the ecosystem is relying on other sources (editorial, competitors, or social) instead of your pages. ### Sentiment (%) Sentiment summarizes how AI describes your brand across dimensions (fees, reliability, security, etc.). It is shown as an overall score plus a dimension breakdown. **Where it appears:** Home KPI; Analytics → Sentiment **How to interpret:** * If sentiment drops, drill into the dimension and read the underlying negative mentions. * Sentiment improves fastest when you address objections with evidence on citeable pages. ### Products: Mentioned / Not mentioned In Analytics → Products, each product is labeled based on whether it appears in AI answers for your tracked prompt set. Product monitoring depends on your prompt set. If you don’t track prompts where products could be recommended (for example “best basketball shoes”), products will often show Not mentioned. ## Platforms and regions ### Platform The AI platform where the run occurred (ChatGPT, Gemini, Perplexity, AI Overviews / AI Mode, Copilot, etc.). Platform matters because different platforms rely on different sources. ### Region / Country The geographic context for the run (if enabled). Region matters because answers and sources can differ by geography. ## Website concepts ### Website → Pages Pages scores and lists your URLs, surfacing technical and content recommendations that improve crawlability and citeability (for example: adding FAQs, adding schema, fixing meta descriptions). ### Website → Referrals Referrals shows click-through sessions from AI platforms as measured in your connected Google Analytics property (and optional Revenue if GA ecommerce is enabled). This is directional and typically undercounts total AI influence. ### Website → Crawlers Crawlers shows AI crawler/bot visits when connected (currently via Vercel). It helps validate discovery and diagnose crawl blockers. ## Action concepts ### Opportunities Opportunities is a prioritized list of content recommendations designed to improve visibility, citations, and rank for your tracked prompt set. ### Content Content is where you create/manage briefs and generate articles: * briefs can come from Opportunities or be created manually, * briefs can be generated into articles, * articles can be refined and then copied into your CMS. ### Off-page Off-page is split into: * **Editorial**: most-cited editorial targets for your prompts * **Social**: most-cited social/community targets for your prompts * **Directories:** most-cited directory targets for your prompts All three lists refresh weekly. The goal is to add trusted language about your brand where AI already trusts sources. ### Brand Kit Brand Kit stores your brand identity, voice/style, writing samples, and a Knowledge Base. It helps generated work stay accurate and consistent. Use Knowledge Base for deeper materials like differentiators, claims boundaries, and competitive context. ## Settings concepts ### Search terms Search terms are the words Meridian uses to detect brand mentions in AI responses. If search terms are too narrow, you can miss mentions; if too broad, you can over-count. ### Seats / roles Members have roles (Member, Admin). Roles control access to settings and management actions. Seats used shows your current seat utilization. For more Answers, visit [answers.trymeridian.com](http://answers.trymeridian.com) # Billing Source: https://docs.trymeridian.com/guides/settings-and-support/settings/billing View or manage your subscription plan and billing settings. Billing is where you manage your workspace subscription settings. This is typically used by admins when updating plan details, confirming your plan type, or coordinating changes with your team. Settings → Billing showing subscription plan and Save ## What you can do here ### View your current plan Billing shows your current subscription type (for example, Enterprise). This is useful for understanding: * what features are included, * and what limits may apply (seats, articles, exports, etc., depending on your plan). ### Update billing settings (when editable) If the plan field is editable in your workspace, you can update it and click **Save**. If your billing is managed externally (common for Enterprise), you may see limited controls here and use **Manage your subscription** from Preferences instead. If you do not see the option you expect (or you cannot edit the plan), it is usually permission-based or plan-managed. Ask a Superadmin, or contact support. ## Common questions ### “Why can’t I change billing?” Billing changes are often restricted to admins, and some plans are managed via a separate subscription portal or contract process. ### “Does billing affect what I can do in the product?” Yes. Plan level may affect limits like seats, exports, or content generation quotas. If you hit a limit, you’ll usually notice it in the relevant feature (e.g., article counter in Content). # Company Profiles Source: https://docs.trymeridian.com/guides/settings-and-support/settings/company-profiles Manage each brand/client profile, its integrations, and the search terms used to detect mentions in AI responses. Company Profiles let you manage multiple brands (or clients) inside one Meridian organization. Each profile can have its own integrations (like Google Analytics or Vercel) and its own **search terms**, which Meridian uses to detect when the brand is mentioned in AI answers. This is where you go when: * you want to connect or disconnect integrations for a specific profile, * you want to update how a profile is named in the dashboard, * or you want to tune mention detection using search terms (brand name variants, product names, etc.). Settings → Company Profiles showing profiles and integrations ## What you can do here ### Rename a company profile Use the profile’s menu (often “…”), then choose edit/rename. You’ll see an edit modal like: Edit company profile modal The name you set here is the display name shown throughout the dashboard. ### Manage integrations for a profile Each profile can show connected integrations (for example: Google Analytics connected, Vercel connected). If you open **Manage integrations**, you’ll see the active integrations for that profile. Typical actions include: * disconnecting an integration for the profile, * or removing it (depending on your UI and permissions). Integrations are profile-specific. If you manage multiple profiles, connecting GA for one profile does not automatically connect it for others. ### Manage search terms (mention detection) Click **Manage search terms** for a profile to open the search terms modal. Manage search terms modal Search terms are the words Meridian uses to find mentions of your brand in AI responses. This commonly includes: * your brand name, * product names, * and common variations or abbreviations. #### How to add good search terms * Include your primary brand name (exact spelling). * Add common variations users might type (spacing, punctuation, abbreviations). * Add flagship product names if AI answers often recommend specific products. * Avoid overly generic terms that could match unrelated content (e.g., “Base” or “Plus” alone). If your search terms are too narrow, you may miss real mentions. If they are too broad, you may over-count mentions that aren’t actually about your brand. ### Disconnect integrations On the Company Profiles list, you may see integration rows with a **Disconnect** action. Use this if you want to stop pulling referral/crawler data for that profile. ## How this connects to the rest of Meridian * Profile integrations power Website tabs like **Referrals** (GA) and **Crawlers** (Vercel). * Search terms influence mention detection, which impacts metrics like Visibility and Prominence. If mention detection looks off, search terms are one of the first things to review. # Exports Source: https://docs.trymeridian.com/guides/settings-and-support/settings/exports View, download, and manage your data exports for a specific company profile. Exports is where you can view and download files generated from Meridian (for example, Prompts exports and Citations exports). It is useful when you want to share results with teammates, send data to an agency, or keep an archive of reporting snapshots. Exports are **company-profile specific**, so the first step is always selecting which profile you want to view. Exports ## Select a company profile Use the dropdown in the top right to choose a profile. Until you select one, you will see an empty state. Once selected, you’ll see a table of exports for that profile. ## What you see in the exports table ### Date range The time window the export covers (for example: Dec 16–Dec 23). This is important when you use exports for reporting, because it tells you exactly which period the data represents. ### Type The export category, such as: * **Citations** * **Prompts** (Your workspace may support additional export types over time.) ### Status Each export has a status so you can tell whether it is ready: * **Completed**: the file is ready to download. * **In progress**: the export is currently being generated. * **Failed**: the export did not generate successfully. If an export is stuck in “In progress” for a long time, it often means the job is still running or the system needs to retry. ## Downloading and managing exports In the **Actions** column you’ll typically see: * a **download** icon for downloading the export file * a **trash** icon for deleting the export entry Use deletion to keep your export list clean (for example, removing failed or outdated exports you no longer need). Exports are a snapshot. If you want the latest numbers, generate a new export for your current date range rather than relying on an older file. ## Common questions ### “Why do I see Failed?” A failed export usually means something went wrong during generation. The most common causes are temporary job issues or missing data in the selected period. If you see failures consistently: * try a wider date range, * confirm the profile has prompts/citations data, * and contact support with the export type and date range. ### “Why do exports depend on company profile?” Profiles have separate prompt sets, integrations, and datasets. Exports follow the same boundaries so data stays clean and shareable. # Manage Organization Source: https://docs.trymeridian.com/guides/settings-and-support/settings/manage-organization Permanently delete your organization and all associated data. Manage Organization contains destructive organization-level actions. This page is primarily used to **delete an organization**. This is intended for cases like: * you created a workspace by mistake, * you are decommissioning a workspace, * or you are cleaning up a test environment. Manage Organization page showing “Delete organization” Deleting an organization is permanent. It removes the organization and associated data from Meridian servers. Only proceed if you are sure you no longer need the workspace. ## Delete organization To delete an organization: 1. Go to **Settings → Manage Organization**. 2. Click **Delete organization**. 3. Confirm the action in the modal by typing the organization name (or the required confirmation text). 4. Click **Delete organization** to finalize. Delete organization confirmation modal This confirmation step exists to prevent accidental deletion. ## Before you delete (recommended checklist) Most teams do these checks first: * Download any exports you want to keep (Settings → Exports). * Confirm which company profiles are in the organization (Settings → Company Profiles). * Confirm whether teammates still need access (Settings → Members). * If you are migrating to a new org, confirm the new workspace is set up and working before deletion. ## Common questions ### “Who can delete an organization?” This is typically restricted to Admins or the highest permission role in the workspace. ### “Can Meridian restore a deleted organization?” Treat this as irreversible unless your contract/support agreement explicitly states otherwise. # Members Source: https://docs.trymeridian.com/guides/settings-and-support/settings/members Manage who has access to your organization, their role, and which company profiles they can access. Members is where you manage user access. This includes inviting new teammates, assigning roles, and controlling which **Company Profiles** each member can access. If your organization manages multiple brands or client profiles, this page is essential for keeping access clean and preventing accidental edits in the wrong profile. Settings → Members showing seats used, roles, and member list ## What you can do here ### Invite a new member Click **Add member** to invite a teammate. Invite new member modal In the invite modal you can set: * **Email address**: the person you’re inviting. * **Role**: Member / Admin / Superadmin (available options depend on your workspace). * **Profile access**: which Company Profiles this person can access. #### Profile access behavior (important) You’ll see guidance like: “Leave empty to grant access to all profiles.” That means: * If you select **no profiles**, the member gets access to **all** company profiles. * If you select specific profiles, the member gets access **only** to those profiles. This is useful if you manage multiple brands and want teammates to work only on certain profiles. ### Change a member’s role In the member list, use the role dropdown (e.g., Admin / Member) to change permissions. Roles typically determine who can: * manage settings, * manage integrations, * invite/remove members, * and edit key workspace configuration. Exact permissions can vary by plan and configuration. If you’re unsure, use the practical rule: Superadmins manage org-wide settings; Members manage day-to-day. ### Edit a member’s Company Profile access If you click the “…” menu or access control on a member, you may see an **Edit profile access** modal. Use this when: * a teammate should only work on one brand/client profile, * or you want to remove access without removing the user entirely. ### Remove a member If you click the trash/remove icon, you’ll see a confirmation modal. Removing a member: * revokes their access to the organization immediately, * and prevents them from accessing any company profiles in that org. ## Seats used At the top of Members you may see “X / Y seats used.” This helps you track capacity and avoid hitting seat limits when inviting new teammates. ## Best practices * Use profile access to prevent “wrong account” mistakes when you manage multiple brands. * Keep Admin access limited to 1–2 owners. * If someone only needs visibility, use a lower role and restrict profile access. # Preferences Source: https://docs.trymeridian.com/guides/settings-and-support/settings/untitled-page Set your organization name and other workspace-wide defaults. Preferences is where you manage basic organization-level settings that show up throughout Meridian. Most teams touch this page when they first set up their workspace (naming) and occasionally when ownership or billing changes. Settings → Preferences showing organization name ### Update your organization name Your organization name is the display name used across the dashboard (for example in headers, settings, and shared views). To update it: 1. Go to **Settings → Preferences**. 2. Edit **Organization name**. 3. Save your changes (if a save action is shown in your UI). If you manage multiple brands inside one organization, your organization name is not the same thing as a “Company Profile.” Company Profiles are configured in **Settings → Company Profiles**. ### Manage subscription If you see a **Manage your subscription** button, it will take you to your subscription management flow (plan changes, billing owner actions, etc.). If you are on an Enterprise plan, subscription changes may be restricted to specific roles. # Citation Rate Source: https://docs.trymeridian.com/guides/start-here/main-metrics/citation-rate Citation Rate is **how often your content is cited** in AI answers. In other words: when an AI assistant includes citations, Citation Rate measures the share of runs where **at least one of those citations points to your owned domain(s)**. **Calculation** $Citation\ Rate = \frac{\#\ responses\ where\ your\ content\ was\ cited}{\#\ total\ responses} \times 100$ ### **Example** If Meridian captured 200 runs in the selected period and 35 of those runs cited your site at least once: * Citation Rate = 35 / 200 = **17.5%** ## **How to interpret it** * A **higher citation rate** usually means AI assistants are relying on your content as a source more often. This typically improves trust and can support better rankings over time. * A **drop in citation rate** often means the assistant is choosing other sources instead of your site for those prompts (competitors, editorial sites, or social/community sources). * If your citation rate is low, the next step is usually: * improving your “source pages” so they are more citeable (\*\*Website → Pages\*\*), * earning more third-party mentions that lead to citations (\*\*Off-page\*\*), * and publishing content aligned to what AI is already citing (\*\*Opportunities / Content\*\*). # Prominence Source: https://docs.trymeridian.com/guides/start-here/main-metrics/prominence **Prominence is your average rank when you are mentioned.** Lower is better. If the assistant returns a list of options, the difference between #1–#2 and #5–#6 is often the difference between being recommended and being "also listed." ### **Calculation** $Prominence=average(position ∣ mentioned)$ ### **Example** If you were mentioned 5 times and your positions were #1, #2, #4, #3, #2: * Prominence = (1 + 2 + 4 + 3 + 2) / 5 = **#2.4** ### **How to interpret it** * Visibility can be stable while prominence worsens. That usually means you are still included, but competitors are being preferred. * Prominence improves when: * Your "why choose us" criteria matches what the answer emphasizes * You have stronger evidence/citations * And your pages are structured in a way models can summarize confidently Prominence is an "average when mentioned," not an overall ranking across the entire market. It is scoped to your tracked prompts and the competitor set that appears in your runs. # Sentiment Source: https://docs.trymeridian.com/guides/start-here/main-metrics/sentiment Sentiment summarizes the “story” AI tells about your brand when answering your tracked prompts. Meridian extracts and scores sentiment across common decision dimensions (for example: fees, reliability, security, customer service, ease of use). Sentiment matters because many prompts are ultimately decision prompts. Even if you are mentioned, negative narrative often acts like a penalty: it can lower your rank, reduce prominence, and make competitors feel like the safer default. ## What Sentiment is Sentiment is not a review score of your company. It is a summary of the language AI uses about you in answers. It helps you answer: * Are we being described as a strong option, or with recurring objections? * What are the exact objections (fees, trust, reliability, support, etc.)? * Are those objections concentrated on one platform or topic? * Are we being described negatively even when we’re recommended? ## How to interpret it If sentiment declines, go to **Analytics → Sentiment** and identify: * Which dimension is driving the decline (fees vs reliability vs security, etc.) * Whether the negatives are concentrated on one platform (for example ChatGPT vs Gemini) * And what the negative language actually says (the exact phrasing is usually actionable) A common pattern is: * Visibility stays stable (you’re still included), * but sentiment worsens, * and prominence drops (you slide down the list). That usually means you’re being included, but competitors are being preferred because the answer frames them as safer, cheaper, more reliable, or easier. ## Typical fixes (what can help) Sentiment improves most often when you add clarity and evidence on pages AI can cite. Common fixes include: * Add clear, comparable pricing information (examples, caveats, what affects pricing). * Add trust and reliability proof (audits, guarantees, SLAs, incident history with remediation). * Add customer support clarity (channels, hours, response expectations, escalation path). * Add FAQs that directly address repeated questions and objections. * Fix the narrative on citeable pages (not only on marketing copy). The goal is not “better adjectives.” The goal is to make the evidence and positioning so clear that AI has no reason to repeat the same objection. # Share of Voice Source: https://docs.trymeridian.com/guides/start-here/main-metrics/share-of-voice Visibility answers: "Are we included?" Share of voice answers: "How much of the conversation do we own vs competitors?" ### **What Share of Voice means** When an AI assistant answers a "best \*\*" or "top \*\*" prompt, it usually returns a short list of options. Share of voice measures how often you appear in those lists relative to the other brands that show up across runs. In practice, share of voice is a good metric for: * tracking whether you are gaining or losing ground against competitors over time, * identifying prompts where competitors dominate the "slots," * prioritizing which prompt themes to invest in (content, off-page, and website fixes). ### **Calculation** For the selected prompt (or prompt set), Meridian counts mentions of each tracked brand across runs.\ \ $SOV = \frac{\#\ mentions\ of\ you}{\#\ mentions\ of\ all\ tracked\ brands} \times 100$ ### **Example** If, across 100 responses for a prompt, the assistant mentions brands a total of 500 times: * You were mentioned 90 times * All tracked brands combined were mentioned 500 times Then: * SOV = 90 / 500 = **18%** This means you occupy about 18% of the overall recommendation space across those runs. ## **How to interpret it** ### **If share of voice is low** Competitors are taking most of the slots across runs. This typically happens when: * competitors are being cited more often (citation ecosystem advantage), * competitors have stronger comparison/list pages that match the prompt format (format advantage), * or competitors have clearer positioning on the criteria the assistant uses (positioning advantage). A low SOV is a strong signal that improving *only one page* may not be enough—you may need a cluster of coverage (content + citations) for that prompt theme. ### **If share of voice is rising** You are appearing more often relative to competitors. This often happens after: * publishing content that matches the prompt format (listicles, comparisons, guides), * improving citeability on your owned pages (FAQ + schema + clearer structure), * earning inclusion on key third-party sources (Off-page Outreach), * or increasing product/category coverage so you are eligible to be listed. A rising SOV is usually a good sign that you are entering the consideration set more consistently and taking share away from competitors. ### **If share of voice is flat but prominence is improving** This pattern means you are not necessarily showing up more often than before, but when you do show up you are ranking higher. That can still be a meaningful win—especially for high-intent prompts where top 1–2 positions drive the most value. ### **If share of voice changes suddenly** Sudden moves often come from: * a small sample size (few runs), * one platform shifting its sources or answer structure, * changes to your competitor set, * or major changes in the citation ecosystem. In these cases, widen the date range and confirm the trend is consistent before acting. # Visibility Source: https://docs.trymeridian.com/guides/start-here/main-metrics/visibility **Visibility tells you how often your brand appears in AI answers** for the prompts you track. ### **Calculation** $\text{Visibility}=\frac{\#\text{ responses where you are mentioned}}{\#\text{ total responses}}\times 100$ ### **Example** If Meridian ran 200 prompt-runs today, each on 4 models, there'd be 800 responses. If you were mentioned in 96 of them: * Visibility = 96 / 800 = **12%** ### **How to interpret it** * A low visibility percentage means you are frequently not included in answers. In practice, this usually means you are **not consistently in the "consideration set."** * Improving visibility tends to require one (or more) of: * Better topic coverage (content that directly answers the prompts) * Stronger presence in the sources models trust (citations) * Or better site discovery (crawlability/citeability). # Improve off-page citations Source: https://docs.trymeridian.com/guides/workflows/improve-citations Earn citations from the sources AI already trusts in your category (editorial, directories, and social). Citations are one of the strongest drivers of stable AI visibility. In many categories, AI assistants build answers by pulling from a relatively small set of trusted sources. If your brand is missing from those sources, you can struggle to show up consistently—even if your website content is strong. This workflow focuses specifically on **off-page** ways to improve citations: * **Editorial** (third-party articles and “best X” lists) * **Directories** (structured listing sources) * **Social** (UGC/community sources like Reddit and forums) If you want to see which sources matter most in your category, start in **Analytics → Citations** first, then execute here. ## Step 1. Identify which off-page sources matter for your prompts Before you take action, you want to make sure you’re working on sources that actually influence the queries you care about. 1. Go to **Analytics → Citations**. 2. Set your date range (Last 7 days vs Prev. Period is a good default). 3. Note whether the biggest drivers of answers are: * Editorial, * Directories, * Social, * or a mix. This tells you where off-page effort is likely to compound. Off-page works best when it’s tied to your prompt set. If you change or expand your prompts, the “most important” sources can change too. ## Step 2. Decide which off-page surface to focus on first Use this decision guide: ### If “best X” lists and publishers dominate → start with Editorial Editorial is best when AI answers cite listicles, reviews, and roundups. ### If structured listing sites dominate → start with Directories Directories are best when your industry relies on structured databases or review-style directories (industry-agnostic or industry-specific). ### If Reddit/forums/community dominate → start with Social Social is best in categories where UGC is heavily cited (often DTC and consumer purchase prompts). You don’t need to do all three every week. Pick one surface, do a small cadence, then measure week over week. ## Step 3. Use Editorial Go to **Off-page → Editorial**. Editorial is a list of third-party articles and domains that are actively being cited by AI for your tracked prompts. It filters down to targets that are most likely to matter for visibility. ### How to work the Editorial list 1. **Start with the highest-impact targets.** Prioritize sources that show up repeatedly for your prompt set. 2. **Open the page and confirm the gap.** Confirm competitors are included and you are not (or your listing is weak/outdated). 3. **Find the right inclusion path.** Look for the author/editor, submission form, editorial guidelines, or update process. 4. **Request inclusion with specific, evidence-backed language.** Focus on: * what you do and who you’re for, * what differentiator matters for the criteria the list is using, * and proof signals (certifications, policies, benchmarks, pricing clarity, etc.). ## Step 4. Use Directories Go to **Off-page → Directories**. Directories are listing sources that AI often treats as structured evidence. Some are industry-agnostic; others are industry-specific (for example, legal directories like [superlawyers.com](http://superlawyers.com)). ### How to work the Directories list 1. **Start with directories that are clearly influencing your category.** If a directory appears repeatedly in citations for your prompt set, it’s usually worth prioritizing early. 2. **Get listed (or strengthen your listing).** Focus on adding language that matches how AI answers rank options: * clear positioning (who you’re for), * concrete differentiators, * proof signals, * “best for” framing. 3. **Avoid vague marketing language.** Directory listings tend to work best when they read like structured facts. The goal is not just a link. It’s to add accurate, differentiated language to a source AI already trusts. ## Step 5 — Use Social (formerly Engage) Go to **Off-page → Social**. Social is a list of community pages (Reddit, forums, LinkedIn, etc.) that are actively being cited by AI for your tracked prompts. UGC importance varies by industry: * some categories rely heavily on community sources, * others barely cite them at all. ### Two ways teams use Social 1. **Participate in existing threads that AI already trusts.** Add helpful, factual language that matches what users are asking. The goal is to shape the narrative and decision criteria in a trusted source. 2. **Use threads as inspiration for your own social content.** Learn what questions repeat, what objections come up, and what “best for” language is repeated—then publish your own posts on the platforms AI is citing. In many cases, links matter less than the language itself. The goal is to get accurate, differentiated language about your brand and products into places AI already trusts. In many cases, links matter less than the language itself. The goal is to get accurate, differentiated language about your brand and products into places AI already trusts. ## Step 6 — Measure impact (week over week) After you do off-page work, measure impact in three places: 1. **Analytics → Citations** * Are the driver sources shifting? * Are you appearing more often in the sources that matter? 2. **Analytics → Prompts** * Did visibility improve on the prompt clusters those sources influence? * Did prominence improve (lower is better)? 3. **Home** * Are citation rate and visibility moving in the right direction? No single off-page action is guaranteed to increase visibility. Outcomes depend on whether the source remains cited for your prompts, whether your mention is aligned with the criteria the answer uses, and whether competitor sources still dominate. Meridian helps you focus on the sources that are most likely to compound. # Improve product recommendations Source: https://docs.trymeridian.com/guides/workflows/improve-product-recommendations Increase how often your products are recommended in AI answers (SKU-level outcomes). This workflow is for product-based brands that want their specific SKUs to show up more often in AI answers, including shopping-style results. Products is most relevant for brands that sell physical products that can show up in AI shopping-style results (product cards). Right now, Products supports Shopify merchants by importing SKUs from Shopify. We’ll add more catalog integrations over time. ## Step 1. Make sure your prompt set includes product/category prompts Products performance depends on your prompt set. If you aren’t tracking prompts where products could be recommended, products will appear as Not mentioned even if you’re doing everything else right. A good default: * Create a Topic per category (e.g., “Basketball shoes”) * Track at least 10 prompts per topic: * “best basketball shoes” * “best shoes for indoor basketball” * “basketball shoes for wide feet” * “basketball shoes vs running shoes” (if relevant) ## Step 2. Start with Not mentioned products Go to **Analytics → Products** and scan for products marked **Not mentioned**. Prioritize: * best sellers, * high-margin products, * or categories where you want to win. ## Step 3. Inspect responses for products that are Mentioned For a product that is Mentioned, click **See responses**. Use the table to learn: * which prompts trigger recommendations, * which competitors show up instead, * what criteria the answer uses (price, fit, durability, trust), * and which sources are cited (if applicable). This tells you the “winning pattern” you should replicate. ## Step 4. Fix product eligibility (structure + criteria coverage) Product improvements that most often increase mentions: * Clear specs and “who this is for” sections * Comparisons against common alternatives * FAQs that match buyer questions * Category/list pages that include your products and match prompt formats If your brand is mentioned but your products are not, the fix is usually product page structure and recommendation criteria coverage—not general brand awareness. ## Step 5. Publish supporting content that matches the prompt format Use **Opportunities** and **Content** to publish: * listicles for “best X” prompts, * comparisons for “X vs Y” prompts, * guides for “how to choose X” prompts. Make sure the content includes: * decision criteria, * “best for” segments, * and concrete product mentions aligned to those criteria. ## Step 6. Measure impact Over the next period: * more products should move from Not mentioned → Mentioned, * prominence should improve on prompts where products appear, * and your owned product/category URLs should begin appearing more often in citations (where citations exist). ## Related * Analytics → Prompts (to confirm prompt coverage) * Website → Pages (to improve product/category citeability) * Publish end-to-end # Improve website readiness Source: https://docs.trymeridian.com/guides/workflows/improve-website-readiness Make your site easier for AI systems to discover, understand, and cite. Website readiness is the foundation layer. If your key pages are not citeable or discoverable, it is difficult to improve owned citations and maintain top prominence—even if you publish a lot of content. This workflow focuses on two product areas: * **Website → Pages** (fix citeability and structure issues) * **Website → Crawlers** (validate discovery via crawler visits, when connected) ## Step 1. Start with Website → Pages Go to **Website → Pages** and scan the top summary indicators (robots health, performance, technical SEO, content). Then prioritize: * pages that matter to your business (pricing, trust, category hubs, key guides), * pages with issues flagged, * template-level issues that affect many pages. ## Step 2. Open a flagged page and apply recommendations Click into a page with issues. You will typically see recommendations like: * add an FAQ section, * add schema, * fix missing meta descriptions, * validate structured data. If Meridian generates copy-ready drafts (FAQ text, schema JSON-LD), click into the recommendation and copy the first draft as your starting point. ## Step 3. Focus on “source pages” first If your goal is to improve citations and prominence, the highest-leverage pages are source pages: * category guides, * pricing/trust pages, * comparison pages, * “best for” pages. These pages should include: * direct answers early, * clear headings, * FAQs (with schema when appropriate), * proof for differentiators. ## Step 4. Validate discovery with Crawlers Go to **Website → Crawlers**. Crawlers currently supports a **Vercel integration**. If you haven’t connected it, you’ll see an empty state prompting you to connect Vercel. Once connected, use Crawlers to confirm: * AI crawler visits are happening, * key pages are being visited, * crawl activity increases after you publish or fix important pages. ## Step 5. Measure impact You will usually see impact in this order: 1. more crawler visits to key pages (leading indicator), 2. more owned citations in Analytics → Citations, 3. improved prominence on prompts tied to those pages. If crawl improves but citations do not, the next bottleneck is usually page structure and evidence, not discovery. # Improving visibility Source: https://docs.trymeridian.com/guides/workflows/improving-visibility A practical framework for increasing how often AI assistants mention, rank, and cite you; then how to execute it inside Meridian. AI visibility is not one thing. Getting recommended consistently usually requires improving multiple “surfaces” of the internet that AI systems rely on. Meridian is designed to help you identify which surfaces matter most for your category and turn that into a repeatable plan. ## The 6 core surfaces that matter for AI visibility Most visibility improvements come from five surfaces. You do not need to work on all five at once, but it helps to understand what each one controls. ### 1) Website backend code (crawlability + machine readability) This is everything that determines whether AI crawlers and retrieval systems can access and parse your pages reliably: * robots.txt and bot protection * sitemaps and internal linking * canonical/noindex behavior * performance and stability * structured data (schema) and metadata If this layer is broken, your content may not be discoverable or citeable even if it is high quality. **How Meridian helps** * **Website → Pages** flags technical issues and generates implementation templates (including schema drafts). * **Website → Crawlers** helps validate that AI crawlers are actually visiting your site (currently via Vercel). ### 2) Existing front-end content (clarity + criteria coverage) This is the content you already have on your site, and whether it answers the questions AI is trying to answer: * clear headings that map to user intent * direct answers early (so content is easy to summarize) * decision criteria and trade-offs * proof and specifics (numbers, policies, comparisons) * FAQs that match real buyer questions Most teams have enough existing content, but it is not written or structured in a way that’s easy for AI to cite. **How Meridian helps** * **Website → Pages** identifies content gaps and generates copy-ready improvements (FAQ drafts, templates, recommended placement). * **Analytics → Sentiment** helps you see which objections you need to address with evidence on the page. ### 3) Net-new content (coverage for prompts you want to win) Net-new content is how you expand coverage into queries where you currently have low visibility: * “best X” prompts * comparisons and alternatives * use-case pages (“X for beginners,” “X for Y need”) * category guides and source pages This is often the fastest way to increase visibility on higher-funnel prompts—especially when you intentionally write content that matches how AI answers are formatted. **How Meridian helps** * **Opportunities** suggests what to create next based on content gaps and citation ecosystems. * **Content** takes an opportunity → creates a brief → generates an article → helps you refine and publish. * **Brand Kit** makes generated content accurate, specific, and on-brand. ### 4) Off-page editorial and “owned” third-party sources (trust transfer) Many categories are shaped by third-party editorial sites and “best X” lists. AI often relies on these sources as evidence. If you are missing from them, you can struggle to rank consistently—even with strong owned content. This includes: * editorial lists and reviews * editorial articles * industry associations * publications and comparison sites This surface is usually about **earning inclusion** rather than publishing on your own domain. **How Meridian helps** * **Off-page → Editorial** is built from the most-cited editorial sources for your prompts, filtered to remove: * sites you’re already included on * competitor domains * Outreach updates weekly so you always see the current citation ecosystem. ### 5) Off-page social / UGC (language and narrative in community spaces) In some industries (often DTC and consumer purchase prompts), AI relies heavily on UGC sources like Reddit threads, forums, and social discussion. In others, social sources barely matter. This surface is less about links and more about **trusted language** existing in places AI already uses as sources: * how people describe your category * what objections come up * what differentiators matter * what “best for” positioning is repeated **How Meridian helps** * **Off-page → Social** surfaces the most-cited social/community pages for your prompt set. * Teams use Social in two ways: 1. participate in existing threads that are actively being trusted by AI for your prompts 2. use those threads as inspiration for what to post next (and where) on your own social channels In many cases, links matter less than the language itself. The goal is to get accurate, differentiated, helpful language about your brand and products into places AI already trusts. ### 6) Directories (structured listing sources) Some industries rely heavily on directories as evidence (industry-agnostic and industry-specific). These often show up in AI answers because directory listings are structured and easy to summarize. **How Meridian helps** * **Off-page → Directories** surfaces the highest-impact directories for your industry that are being cited for your prompts, where you are not currently included (or not being cited). ## Where review websites fit Review sites often sit in between “editorial” and “UGC,” and they can be extremely influential in categories where buying decisions are comparison-heavy. Examples include: * “best X” review publishers * category-specific review sites * marketplaces and directories * community review platforms If your category’s citation ecosystem is review-heavy, they often show up prominently in: * **Analytics → Citations** (right panel / driver sources) * **Off-page → Outreach** targets **How to treat review sites** * If the site is actively cited for your prompts, prioritize inclusion and accuracy there. * Focus on being described with the criteria that matter for your prompts (not generic hype). * Provide a strong, citeable source page on your site that supports any claims. ## How to use Meridian to improve visibility (end-to-end) This is a practical sequence that works well for most teams. ### Step 1 — Make sure you’re tracking the right prompts Go to **Analytics → Prompts**. Prompts should represent the queries you want to show up more for in AI search. It is normal (and valuable) to track higher-funnel prompts where you currently have low visibility. Avoid building a prompt set that is mostly branded or too company-specific, because it can inflate visibility without measuring discovery. ### Step 2 — Diagnose one important prompt In **Analytics → Prompts**, pick one “money prompt” and open **Analyze**. Inspect: * who outranked you, * what criteria the answer used, * and what sources were cited. This tells you which surface is the bottleneck: * site readiness, * content structure, * net-new coverage, * editorial inclusion, * or social/UGC narrative. ### Step 3 — Use Citations to identify the evidence ecosystem Go to **Analytics → Citations**. Use it to understand: * which sources AI trusts for your prompt set (the corpus you must compete with), * and where your domain appears (or is missing). This usually determines whether you should prioritize: * Website → Pages fixes, * Off-page Outreach/Engage, * or Content creation. ### Step 4 — Execute on the right surface #### If the bottleneck is website readiness Go to **Website → Pages** and fix high-leverage source pages: * add FAQs and schema, * improve clarity and structure, * fix metadata and structured data issues, * use Meridian’s copy-ready drafts where available. If Pages is empty, allow-list MeridianBot so we can crawl your site. #### If the bottleneck is coverage (you’re not mentioned) Go to **Opportunities** and generate a brief. Then go to **Content** and publish a page that matches the prompt format: * listicle for “best X” * comparison for “X vs Y” * guide for “how to choose X” #### If the bottleneck is citations (you’re not in the sources) Go to **Off-page → Outreach**. Pick the highest-cited editorial targets for your prompt set and pursue inclusion. If social is a major driver in your category, use **Off-page → Engage** to participate in the threads AI is already citing. #### If the bottleneck is narrative (you’re mentioned but not preferred) Go to **Analytics → Sentiment**. Find the dimension that’s driving negativity (fees, trust, reliability, support). Then fix it where AI can cite it: * update source pages (Website → Pages) * add proof and clarity * influence off-page sources if the narrative is driven externally ### Step 5 — Measure impact week over week Visibility improvements are rarely immediate. Measure over a full period (usually 7 days) and compare to Prev. Period. Watch: * **Visibility** and **Prominence** on target prompts (Analytics → Prompts) * **Owned citations** increasing (Analytics → Citations) * **Narrative improving** in key dimensions (Analytics → Sentiment) Adding your name to the internet does not guarantee visibility increases. Many factors influence outcomes: prompt coverage, citation ecosystems, page structure, and platform-specific behavior. Meridian helps you identify what matters most and focus effort where it’s likely to compound. ## A simple starting plan (one week) If you want a safe default plan: 1. Pick 5–10 target prompts. 2. Fix 1–2 key source pages in Website → Pages (FAQ + schema is a strong starting point). 3. Publish 1 net-new piece from Opportunities/Content aligned to those prompts. 4. Do a small Outreach or Engage cadence (depending on what Citations shows is driving your category). 5. Measure next week and repeat. # Publish end-to-end Source: https://docs.trymeridian.com/guides/workflows/untitled-page Go from opportunity → brief → article → publish → measure (repeatable content loop). This workflow is for teams who want a repeatable way to publish content that improves AI visibility over time. It uses Meridian’s built-in recommendations, brief generation, and article workflow, then closes the loop by measuring impact in Analytics. ## Step 0. Set Brand Kit first (so content is accurate) Before generating briefs and articles, fill out **Brand Kit** basics: * Brand mission and ICP * Voice & Style * Writing samples (1–3 is enough to start) * Knowledge Base entries for differentiators, claims boundaries, and competitive context This reduces generic output and prevents risky claims. ## Step 1. Pick an opportunity (recommended) or create a brief directly Option A (recommended): * Go to **Opportunities** * Choose a High priority item * Click **Generate brief** Option B (manual): * Go to **Content** * Click **New brief** * Choose a content type (listicle, comparison, guide, etc.) ## Step 2. Review the brief (don’t skip this) Open the brief in **Content** and check: ### Target Prompts Confirm the brief matches the prompts you care about. If the brief is aimed at irrelevant prompts, it won’t move the metrics you’re tracking. ### Citations guidance Check which sources Meridian believes matter. This often tells you what evidence ecosystem you’re competing in. ### Structure Confirm the outline matches the prompt intent: * “best X” prompts → listicle + criteria + best-for sections * “X vs Y” prompts → comparison table + trade-offs * “how to” prompts → step-by-step + troubleshooting ## Step 3. Generate the article Click **Generate article** and confirm the generation. After it generates: * edit for factual accuracy and product details, * add concrete examples and criteria, * ensure you’re not making claims you can’t support. ## Step 4. Refine (short, specific instructions work best) Use the refinement input for targeted improvements like: * “Rewrite the intro to answer the question in 2 sentences.” * “Add a comparison table with these criteria: …” * “Add 8 FAQs that match buyer questions for this topic.” * “Add 3 competitor alternatives and explain who each is best for.” ## Step 5. Publish on a stable URL Publish the final content to your CMS. If the content should become a “source page,” it should include: * clear headings, * direct answers early, * FAQs and schema where appropriate, * and proof sections that support your differentiators. ## Step 6. Measure impact Over the next period, check: * **Analytics → Prompts** for the target prompts (visibility score, prominence). * **Analytics → Citations** to see whether your domain/URL appears more often. * **Home** KPI deltas (directional trend). If nothing moves after a full period, the most common reasons are: (1) prompt mismatch, (2) competitor/editorial sources dominate citations, or (3) your pages aren’t citeable yet. # Home Source: https://docs.trymeridian.com/home Your scoreboard for visibility, citations, sentiment, and rank. home Home is where most teams start each week. It summarizes whether AI assistants are recommending you more often, ranking you higher, and relying on sources that support you. It also helps you quickly decide what to investigate next. ## Time range and previous period Home is scoped by the date range you select (for example, “Last 7 days”) and the “Prev. Period” comparison. Prev. Period compares to the immediately previous time span of the same length. Example: Last 7 days vs the 7 days right before that. ## KPI cards Home shows four KPIs: * **Visibility**: how often you are mentioned across runs. * **Citation Rate**: how often answers include sources. * **Sentiment**: how positively AI describes you across key dimensions. * **Prominence**: your average rank when you are mentioned (lower is better). If you want the exact formulas, see **Main Metrics.** ## Visibility chart The Visibility chart shows daily visibility for you and competitors. It is useful for spotting: * When a change began * Whether it is consistent or a one-day spike * And whether competitor gains align with your decline. ## Brand Rankings Brand Rankings shows a ranked list of brands and their visibility. This is the fastest way to answer “who is beating us overall?” without picking a single prompt. ## Rising Visibility Rising Visibility helps you see what is improving the most. Use it to: * Double down on topics/prompts that are already trending in your favor * Spot new areas where you may want to add prompts or content * Validate that recent work is moving the right direction. ## Platform Breakdown Platform Breakdown shows how visibility differs across AI platforms. Differences are normal because platforms rely on different sources and retrieval behavior. If one platform is consistently weaker, you usually need a platform-specific plan (especially around citations). # Opportunities Source: https://docs.trymeridian.com/opportunities A prioritized list of content opportunities designed to improve visibility, citations, and rank in AI answers. Opportunities helps you decide what to publish next without guessing. Each opportunity is generated from your performance and the citation ecosystem in your category. The goal is to publish content that AI assistants can reliably summarize, cite, and recommend. Opportunities cards with priority and actions ## How Opportunities are generated Meridian looks at patterns across your tracked prompts and answers, including: * Prompt categories where you are frequently not mentioned * Prompts where you are included but outranked * Citation drivers (which domains and URLs models trust), * And content formats that commonly win the prompt type (listicle, comparison, how-to, guide). Opportunities are not random blog ideas. They are meant to map to specific gaps in what AI is currently citing and recommending for the prompts you want to increase visibility for. ## Actions you can take ### Generate brief Generates a structured brief designed for AI visibility. More information on Briefs in **Content**. ### Decline Decline removes irrelevant opportunities so the list stays useful. Declining is helpful when a suggestion is off-topic for your business or not aligned to your audience. It keeps your backlog clean. ## A practical way to use Opportunities each week 1. Start with High priority opportunities. 2. Choose one opportunity tied to your most valuable prompt themes. 3. Generate a brief and publish a single strong piece rather than many small posts. 4. Measure impact in Home and in Analytics → Prompts for the related prompts. # Quickstart guide Source: https://docs.trymeridian.com/start-here/quickstart Get Meridian set up in 20 minutes, establish a baseline, and take your first actions to improve AI visibility. This quickstart walks you through the fastest path to value in Meridian: 1. Track the prompts that matter 2. Understand your visibility baseline 3. Generate your first brief to start improving visibility. ## 1) Add your first prompts Prompts are the questions you want to win in AI assistants. The quality of your prompts determines how useful your dashboards will be. Go to **Analytics → Prompts**, review the initial prompts generated, and add **10–25 prompts** to start. Prompts ### What good prompts look like Use conversational questions that match how people ask AI tools for recommendations: * “best \_*\_ for \_*\_” * “top \_\_\_ tools/platforms” * “\_*\_ vs \_*\_” * “\_\_\_ for beginners” * “how do I choose \_\_\_” * “is \_\_\_ worth it” The best way to think about your prompt list is: These are the types of queries you want to appear more for in AI search. From here we'll start tracking and improving on them. Add queries across top, middle, and bottom of funnel. ### What to avoid (for now) * Purely branded prompts only (e.g., “What is trymeridian.com?”). These are helpful later, but they won’t tell you whether you’re being discovered. * Prompts extremely specific to your brand (e.g., "What is the best GEO platform based in New York that starts with the letter 'M')\ Adding prompts that you already have high visibility means we'll only be trying to increase visibility in queries that you already appear for. * Prompts that are too broad (e.g., “crypto” or “marketing”). Broad prompts are hard to act on because the answer criteria varies a lot. A simple way to build your first prompt set: take your top SEO pages or categories and rewrite them as questions a buyer would ask an AI assistant. ## 2) Set your competitor context Meridian is comparative by nature because AI answers often return lists. Adding competitors helps you understand who is being recommended alongside you (or instead of you). Go to **Analytics → Competitors** and add **3–8 true substitutes**. Competitors ### Tips for competitor selection * Pick companies your customers would genuinely consider instead of you. * You can always adjust this later as you learn who shows up in answers. ## 3) Check Home to establish your baseline Analytics Home gives you a quick read on: * **Visibility**: how often you’re mentioned across your tracked prompts * **Citation Rate**: how often answers include sources * **Sentiment**: how AI describes you across decision-driving dimensions * **Prominence**: your average rank when you’re mentioned We add data every day. Wait for a few days of data to really understand your baseline. As more data is populated, use “Prev. Period” to compare against the previous time window of the same length. This is the simplest way to see whether you’re trending in the right direction. ### What to do if your baseline looks low That’s the best starting point. We want to add prompts where you want to increase visibility in, instead of solely prompts you're already showing up for. Low baseline usually means one of: * You’re not included in answers yet (coverage/authority gap), * Competitors have stronger source presence (citation gap), * Or your website/content isn’t structured to be cited (site readiness gap). You’ll address these in the next steps. ## 4) Investigate one important prompt Go to **Analytics → Prompts** and pick a prompt that matters to your business. Then: 1. Open the prompt’s analysis (press **Analyze**) 2. Find a run where you were **not mentioned** or ranked low 3. Open the run details to see: * Who outranked you, * What criteria the answer used, * And which sources were cited Analyze(single) This gives you immediate clarity on what is driving your current rank: sources, structure, narrative, or product coverage. ## 5) Add your Brand Kit (so outputs are accurate and on-brand) Go to **Brand Kit** and fill out the basics before you generate content. Brandkit Brand Kit is the “ground truth” Meridian uses when it creates briefs and articles. If it’s empty or vague, generated content tends to be generic. If it’s specific, your content is more likely to reflect the language and criteria AI uses to rank answers (and less likely to include risky or inaccurate claims). At minimum, add: 1. **Positioning (1–2 sentences)** Who you are, who you’re for, and what you do. 2. **Key differentiators (with proof where possible)** List 5–10 differentiators that matter for “best X” prompts (price, trust, support, speed, etc.). Include links to source pages if you have them. 3. **Approved vs disallowed claims** Write down what content can safely say and what it must not say. This is especially important in regulated or high-stakes categories. 4. **Voice & style guidelines** A few rules like “short sentences,” “use bullets,” “include decision criteria tables,” “avoid vague superlatives.” 5. **(Optional) Writing samples** Add 1–3 samples so Meridian can match your tone. ## 6) Generate your first brief (and optionally an article) Go to **Opportunities** and pick one “High” priority recommendation. Opportunities Click **Generate brief**. Then go to **Content** and open the brief. From there you can: * Review the structure and target intent * Generate an article draft * Refine it * And publish it in your website Brief ## 7) Improve your website readiness (fast win) Go to **Website → Pages**. This is where Meridian scores your site for AI discovery and citation-readiness. Pages Start by scanning the top summary scores (Robots, Performance, Technical SEO, Content). These scores help you quickly see what’s limiting visibility; whether AI can crawl you, understand you, and confidently cite you. Next, scroll the page list and a page with an issues flag. Prioritize: * A page that matters to your business (product/category/guide) * A page with a higher issue count or lower Technical/Content score. Click into that page to open its recommendations. You’ll see specific fixes (for example: **add an FAQ section**, **add FAQ schema**, **fix missing meta description**, or other technical/content improvements). Pages Recommendations From there you can: * Review the recommended fix and why it matters * Follow the step-by-step action steps Meridian provides * Implement the change on your website (in your CMS or codebase) If you only make one website change this week, start with a recommendation that improves citeability (like **FAQ + schema**) on a page that should be a “source page” for your category. This is one of the most reliable ways to increase owned citations over time and improve rankings. ## 8) Capture quick wins with Off-page (Editorial, Directories, Social) Go to **Off-page** and pick one surface to work on this week. Off-page is where you earn mentions in the sources AI already trusts for your prompts. A good default is to spend 15–30 minutes and pick a small set of targets you can actually complete. ### Editorial (lists and articles) Go to **Off-page → Editorial** and pick 3–5 high-impact targets. Editorial is a list of third-party articles that are actively being cited for your tracked prompts. If competitors are included and you are missing, requesting inclusion can be a fast way to improve the source ecosystem AI pulls from. When you reach out, include: * who you are and who you’re for, * why you belong in the list, * what differentiator matches the criteria, * and a credible source page that supports the claim. ### Directories (structured listing sources) Go to **Off-page → Directories** and pick 3–5 directory targets. Directories can be high leverage because listings are structured and easy for AI to summarize. When you create or update a listing, focus on adding specific, on-brand language: * what you do and who you’re for, * your differentiators (with proof where possible), * and “best for” positioning that matches how people choose in your category. ### Social (UGC and community) Offpage Go to **Off-page → Social** and pick 2–3 threads or community pages. In some industries (often DTC), AI relies heavily on community sources. Social is less about links and more about trusted language showing up in places AI already cites. Teams use Social in two ways: 1. participate in existing threads that AI is actively citing for your prompts, and 2. use those threads as inspiration for what to post next (and where) on your own social channels. Off-page targets update weekly based on your prompt set. If you change or expand your prompts, your best editorial/directories/social targets can change too. ## 9. Import your products for SKU-level visibility (For Ecommerce only) If you sell physical products that can appear in shopping-style AI results (product cards), you can import your catalog so Meridian can track product-level outcomes. Go to **Analytics → Products** and connect your store to import SKUs (currently Shopify). Products Once your catalog is connected: * Products shows which SKUs are being mentioned vs not mentioned for your tracked prompts. * You can click **See responses** to understand which prompts trigger recommendations, what criteria the answer uses, and which competitors appear instead. Products depends on your prompt set. If you don’t track category/product prompts (for example “best basketball shoes”), you will almost certainly see many products marked **Not mentioned**. A good setup is to create a Topic per category and track \~10 prompts per topic. # Understanding your data Source: https://docs.trymeridian.com/start-here/understanding-your-data APIs vs Scraping UI, what a “response” is, why answers vary, and how to interpret trends correctly. Meridian runs your tracked prompts across supported AI platforms and regions on a recurring cadence, then aggregates patterns over time. This is important because AI answers naturally vary day to day. Analyze Grouped Each prompt run is recorded with: * The prompt text * The platform and region * Whether you were mentioned * Your position/rank when mentioned * The response (or a response snippet) * Citations (domains/URLs), when available * Sentiment signals extracted from the response From there, Meridian turns run-level evidence into dashboards (Home, Analytics, Website) and action queues (Opportunities, Content, Off-page).\ \ AI answers change. Sometimes they change because the model changed. Sometimes they change because the sources changed. Sometimes they change because the prompt is ambiguous. Meridian is built to make that variability usable, so you can measure performance reliably and improve it over time. This page explains what Meridian records, why the same prompt can look different day to day, and how to use the data without overreacting to noise. ## What Meridian runs (and what a "response" means) Meridian tracks your prompts over time across AI platforms (and often regions). Each time a prompt is executed, Meridian records a **response**. Response A response is defined by: * the prompt text, * the platform (ChatGPT, Gemini, Perplexity, AI Overviews / AI Mode, etc.), * the region, * the response from the AI assistant, * and the date/time it was run. ## What Meridian captures from a response From each response, Meridian captures evidence that you can inspect later, including: * **Mentioned vs not mentioned**: whether your brand appeared in the answer. * **Position / ranking**: where you appeared if the answer is list-like (#1, #2, etc.). * **The answer text**: the response (or a snippet), so you can see exactly what was said. * **Citations**: domains and URLs referenced, when the platform provides them. * **Sentiment signals**: narrative cues extracted from the answer across dimensions (fees, reliability, security, etc.). This evidence is what powers Analytics tabs and the actions in Opportunities and Off-page. ## Why the same prompt can change day to day It is normal for answers to vary across days. Common reasons include: ### The platform chooses different sources If the model is retrieval-based, it may cite different URLs on different days. That can change which brands are included and who ranks first. ### The platform changes model behavior AI platforms ship updates frequently. Output style and ranking logic can shift. ### The prompt has multiple valid interpretations For example, "best crypto exchange" could be interpreted by fees, security, geography, or product features. Slight differences can change the ranked list. ### The ecosystem changes Competitors publish new content, new editorial pages appear, and community conversations shift. Those changes can affect citations and rankings. One run is best used for investigation. A week of runs is best used for decisions. Meridian is designed to emphasize the second. ## Technical approach: API access vs UI scraping AI platforms often behave differently depending on how you access them. Some tools provide APIs, but API outputs can differ from what real users see in the product UI (and citation behavior can differ as well). In addition, some platforms do not offer stable public APIs for the exact experience users interact with. Meridian's goal is simple: **measure what your customers see** as closely as possible, consistently, and at scale. ### Why Meridian doesn't rely on APIs alone Using only APIs can introduce gaps such as: * **Different outputs**: API responses can differ from UI responses in tone, structure, and ranking behavior. * **Different sourcing**: Citations and retrieval can vary between API and UI experiences. * **Limited coverage**: Some platforms or product experiences are not fully represented through public APIs. ### How Meridian keeps data consistent Depending on the platform, Meridian uses methods designed to reflect a consistent "typical user" experience and capture results in a repeatable way. The goal is to reduce noise, reflect realistic outputs, and make trends meaningful when you compare weeks. \> In practice: you should treat a single day as a snapshot, and treat week-over-week patterns as the best signal for decisions. ## How Meridian turns variability into useful signals Meridian makes daily variation useful by aggregating runs into: ### Home KPIs (weekly scoreboard) * Visibility * Citation Rate * Sentiment * Prominence These help you track trends and compare periods (Last 7 days vs Prev. Period). Home ### Analytics tabs (drivers) * Prompts: which prompts you win/lose and how that changes * Sentiment: which narratives help/hurt you * Citations: which sources drive answers and where you are missing * Competitors: who is winning and on what prompts/topics * Products: which products are recommended or missing Prompts ### Action surfaces * Opportunities: what to create next based on gaps * Content: briefs and articles designed to improve performance * Off-page: editorial/community targets where inclusion can improve trust Opportunities ## How to interpret "spikes" and "drops" A practical approach: * If something changes on **one day only**, treat it as a signal to inspect (open that run), but avoid large strategic changes. * If something changes **consistently across a few weeks**, treat it as a trend and choose an action. If the change is large and sudden: * check whether filters/time range changed, * check whether the platform breakdown suggests it was one platform, * then inspect the citations used in losing runs. ## How to validate whether an improvement "worked" Improvements typically show up as: 1. You are mentioned more often on the target prompts (visibility up for those prompts). 2. You move higher in the list (prominence improves). 3. Your owned pages appear more often as citations (owned share increases). 4. Negative narratives decrease in the relevant sentiment dimension. For the cleanest measurement, pick a small set of "target prompts" before making changes. Then track those prompts for one or two periods after publishing/fixes. Visibility changes aren't instantaneours. Improvements can take weeks to months to appear, and no single improvement is guaranteed to affect visibility. # Welcome to Meridian Source: https://docs.trymeridian.com/start-here/welcome Meridian helps you measure and improve your visibility in AI search.