Skip to main content
Referrals shows click-through traffic from AI platforms (ChatGPT, Perplexity, Copilot, Grok, etc.) as measured in your connected Google Analytics (GA4) property. Referrals This is useful for two main reasons:
  1. It gives you a directional trend of whether AI-driven discovery is increasing over time (especially after you publish content, earn citations, or improve rankings).
  2. If you have ecommerce revenue tracked in GA, it can show revenue from AI-referred sessions (also directional).
Referrals is best used as a trend monitor, not as a complete ROI number. It only counts users who click directly from an AI platform to your site (and, for revenue, only those who convert within the tracked session).
Referrals uses your Google Analytics integration. If you haven’t selected a property yet, you’ll see an empty state prompting you to link/select one.
  1. Go to Website → Referrals
  2. Click Select property
  3. Choose the GA4 property you want Meridian to use
Once linked, Meridian will start showing AI-platform sessions and (if available) revenue.

​
Referrals view (click-through sessions)

The Referrals view shows sessions where GA attributes the session source to an AI platform domain (e.g., ChatGPT, Perplexity, CoPilot, etc.).

​
What you see

  • Total referrals for the selected time range
  • Breakdown by platform (ChatGPT, Perplexity, Copilot, Grok, etc.) with period-over-period deltas
  • A trend chart so you can see changes over time
  • Recent Visits, including landing page, device, country, and engagement (helpful for debugging and optimization)

​
How teams use this

  • Identify which AI platforms are driving the most click-through traffic.
  • Identify the landing pages AI users hit most often, then improve those pages (clear value prop, FAQs, comparison criteria, internal linking).
  • Use week-over-week changes to validate whether your AI visibility work is translating into more inbound sessions.

​
Practical example

If you publish a “Best basketball shoes” guide and then see:
  • more ChatGPT referrals over the next week,
  • and more sessions landing on that guide URL,
that is a strong directional signal the work is paying off.

​
Important nuance: this undercounts AI influence

Referrals is click-through tracking, not “total influence.” It only counts sessions where someone:
  1. Sees a recommendation in an AI platform, and then
  2. Clicks through to your site in that moment (so GA records the referrer domain like chatgpt.com or perplexity.ai).
A lot of AI-influenced journeys will not show up here, for example:
  • The user gets a recommendation in AI, then Google searches your brand and clicks a search result.
  • The user reads the AI answer, leaves, and comes back later via direct, a bookmark, or another device.
  • The user shares your brand with someone else, who then visits via a different channel.

​
How big is the gap (roughly)?

Public research strongly suggests direct clicks are a small fraction of “people who saw an AI answer.” Practical takeaway: It is normal for GA to capture only a small fraction of AI exposure. Use Referrals to monitor directional trend (up/down by platform and by landing page), not as a complete ROI number.
A helpful mental model: “Referrals = minimum measurable click-through.”It is a lower bound on AI impact, because it only counts immediate clicks that preserve the AI referrer.

​
Revenue view (optional)

If your GA4 property tracks revenue and ecommerce events, you can switch to the Revenue view. Revenue view showing revenue from AI search and platform breakdown

​
What you see

  • Total Revenue, Orders, AOV, and Conversion Rate for the selected period
  • Revenue from AI search (revenue attributed to sessions referred from AI platforms)
  • A trend chart and Revenue by Platform table (which platforms are driving the tracked revenue)

​
When this is most useful

  • You want directional evidence that AI referrals are not just visits, but also producing conversions.
  • You want to compare which platforms drive higher conversion rates or AOV.
This “AI revenue” is a lower bound. It captures only conversions that happen in the measured click-through session (or attribution window configured in GA), and it will miss many “view-through” journeys where AI influenced the buyer but GA credits another channel later.