How to Measure AEO: Tracking Your Visibility in ChatGPT, Perplexity and Google AI

The question we get most often from clients running an AEO programme is the same one: how do I know if it is actually working? Google has Search Console. Paid media has conversion tracking. AEO has no single dashboard. The answer is a combination of three approaches, two of which are free and start paying off from month one.

This guide walks through exactly what to track, how to track it, and what the numbers typically look like over time. It is written for a small business that wants to measure AEO seriously without buying a £500-a-month tool.

Why AEO measurement is genuinely harder than SEO

Traditional SEO measurement works because Google exposes an API (Search Console) that tells you exactly which queries surfaced your site and how many clicks you received. AI engines do not do this. ChatGPT does not report on whether it cited you. Perplexity does not send a report. Google AI Overviews appear in Search Console data but are lumped in with normal results, so you cannot isolate their impact without extra work.

The consequence is that AEO measurement is indirect. You infer citation patterns by sampling (running prompts), detect AI traffic by referral source (when someone clicks from within an AI answer), and track authority by mention counts (which lead AI citations rather than reflect them). None of these are perfect signals. Combined, they are enough to tell whether an AEO programme is working.

The three things worth tracking

• Citation rate. What share of your target queries surface your business in ChatGPT, Perplexity or Google AI Overviews? Measured by manual prompt checks.
• AI-sourced traffic. How many people are clicking through from AI answers to your site? Measured in GA4 via referral source.
• Entity mention growth. How often is your business named on third-party credible sources over time? Measured via Google Alerts, brand monitoring tools, or a simple monthly manual sweep.

Tracking all three together gives a reliable picture. Citation rate tells you about visibility in the moment. Traffic tells you about commercial value. Mentions tell you about the underlying authority signals that drive future citations.

Method 1: Manual prompt checking (free, do this first)

The single most useful AEO measurement costs nothing and takes 30 minutes a month.

Build a list of 10 to 20 queries that matter to your business. Mix high-intent commercial queries ("best carpet cleaner newcastle") with informational queries ("how to remove carpet stains"). Every month, run each query in ChatGPT, Perplexity and a Google search (to catch the AI Overview). Record whether your business is named or cited, and in what position. A simple spreadsheet with four columns (query, ChatGPT, Perplexity, Google AIO) plus a citation count per month is all you need.

Important quirks to know:

• ChatGPT's answers vary between runs. Run each query two or three times and record the most common outcome.
• Use incognito or clear-history sessions so personalisation does not bias the result.
• Perplexity provides its citation list inline, which makes auditing easy. ChatGPT with search on does the same. Google AI Overviews show source cards.
• Local queries behave differently depending on whether you set a UK location. Keep the location fixed to avoid month-to-month inconsistency.

Method 2: GA4 referral tracking (free, second priority)

When someone clicks a link inside a ChatGPT answer or a Perplexity citation, GA4 records that session as referral traffic with a specific source. You do not need to configure anything special. You just need to know where to look.

In GA4, go to Reports, then Acquisition, then Traffic acquisition. Group by Session source. The sources to watch for:

• chat.openai.com / chatgpt.com (ChatGPT)
• perplexity.ai (Perplexity)
• gemini.google.com (Google Gemini)
• copilot.microsoft.com / bing.com/chat (Microsoft Copilot)
• claude.ai (Anthropic's Claude, when search is enabled)

Create a custom exploration report that filters for any session with one of these referral sources. Save it and check it monthly. Two caveats: Google AI Overview clicks are lumped into "google / organic" and cannot be isolated in GA4. Users who read the answer without clicking never show up anywhere. This is why Method 1 is more important than Method 2.

Method 3: Paid tools (when free methods are no longer enough)

For a small business, the free methods are genuinely enough for the first 6 to 12 months. Once your AEO programme matures (you are being cited on enough queries that manual checking gets slow), paid tools become worth the cost. Three categories are worth knowing:

• Dedicated AEO platforms (e.g. Profound, Athena HQ, Peec AI, Otterly). These run scheduled prompt checks across all major AI engines and produce dashboards. Usable from around £150 per month.
• AI mention APIs (e.g. DataForSEO LLM mentions). Lower-level access that lets you build your own measurement. Much cheaper per query but requires developer effort.
• SEO tools adding AEO features (Ahrefs Brand Radar, Semrush AI tools). Useful if you are already paying for the platform; less compelling as a standalone purchase.

A simple monthly AEO report

One page. Five numbers. Compare month on month.

• Citation rate across your 10-20 priority queries (X / 20 for each engine).
• AI referral sessions (total, plus split by source).
• AI referral conversions (did any of those sessions complete an enquiry).
• New entity mentions this month (from Google Alerts / brand monitoring).
• One qualitative note: what notable change appeared this month in any AI engine's answer.

That is the whole report. Anyone selling you a 30-page AEO dashboard as a requirement is selling the tool, not the outcome.

What the numbers typically look like over time

For a small business starting AEO from scratch, the realistic pattern we see across clients:

• Month 1-2: citation rate near zero. AI referral traffic likely zero or single digits. This is normal. AI citations are lagging indicators.
• Month 3-5: first citations appear, usually in Perplexity and Google AI Overviews (both pull fresh web content). ChatGPT is slower because a proportion of its answers come from training data, not live search.
• Month 6-9: citation rate reaches a meaningful share of priority queries. AI referral traffic becomes a tracked channel rather than a curiosity. Entity mentions compound.
• Month 9-12+: stable and growing presence across all three engines. AI referral conversions start to show up as an attributable revenue channel.

If you see nothing after 6 months, the underlying AEO programme probably has a problem, not the measurement. Usually either the content is not yet answer-first, the entity data is inconsistent, or the authority signals have not started to propagate.

How we track AEO for clients

Measurement is built into every AEO and SEO retainer we run. Monthly reports cover Google rankings, AI citation rates, AI referral traffic, and entity mention growth. We also share the raw spreadsheets so clients can see exactly what moved. If you want to see what your business currently looks like to ChatGPT, Perplexity and Google AI, book a free audit and we will walk you through the baseline. For the strategic framework behind the numbers, see our AEO strategy playbook.

Common questions

Can I see AI Overview impressions in Google Search Console?

Partly. Search Console shows impressions and clicks that include AI Overview surfaces, but does not separate them out. You can infer AIO presence by looking for queries where impressions are stable but click-through rate has dropped sharply. Google has announced plans to expose AIO data more directly, but as of early 2026 it is not available.

Does ChatGPT referral traffic show up as direct traffic in GA4?

Only sometimes. When a user clicks a source link inside a ChatGPT answer, GA4 typically records chat.openai.com or chatgpt.com as the referrer. When a user reads the URL in ChatGPT and types it into their browser, or copies it, the referrer is lost and the session records as direct. So direct traffic tends to increase alongside AI presence, and is a soft secondary indicator.

How often should I run the manual prompt check?

Monthly is the right cadence. Weekly adds noise (ChatGPT outputs vary run to run). Quarterly misses movement. Run the same query list in the same way, in roughly the same part of the month, and compare scores.

How many queries should I track?

Between 10 and 25 is the sweet spot for a small business. Fewer than 10 and random variation masks the signal. More than 25 and the manual check becomes a 2-hour job no one will actually do each month. Pick queries that reflect real customer behaviour, not vanity head terms.

Is it worth paying for an AEO tool in the first year?

For most small businesses, no. The free methods capture the movement that matters in the first 6 to 12 months. Once citation rate crosses roughly 25% of priority queries and AI referral traffic is a meaningful channel, paid tools start to earn their cost by scaling the measurement. Before then, spend the money on the content and authority work, not on dashboards.

Editor's note (April 2026): For the technical-readiness side of AI visibility (whether agents can read your site at all), Cloudflare's free Agent Readiness Score is the new benchmark.