To track AEO, you watch two things: whether AI answer engines like ChatGPT and Perplexity mention your brand, and whether those mentions send anyone to your store. The problem is that most guides only measure clicks, and most tools only measure mentions. So you end up with half the picture and a false sense of progress.
AEO, or answer engine optimization, is the work of getting cited when buyers ask AI a question instead of Googling. This guide shows you how to track both halves properly: the metrics that matter, a free GA4 setup, and the tools worth paying for.
What does it mean to track AEO?
Tracking AEO means measuring two separate layers, because AI search splits your funnel in two. First there's visibility, which is whether the answer engine names you when a buyer asks a question. Then there's traffic, which is whether that mention actually sends someone to your site.
These layers need different tools, and mixing them up is the mistake we see most often.
- Visibility (the answer): Are you mentioned or cited inside ChatGPT, Perplexity, Gemini, and Google's AI Overviews? This lives inside the AI response, so your analytics will never show it.
- Traffic (the click): When you are cited, does anyone click through? This shows up in your analytics as AI referral traffic.
One more thing worth clearing up. This is not the same as SEO rank tracking, which tells you where you sit on a results page. AEO tracking asks a different question entirely: are you inside the answer at all? It's the feedback loop for how to optimize for AEO.
- From our work: We run AEO for a lighting brand, and when we started, ChatGPT named three competitors and never them. Nothing in Google Search Console would have flagged that. You only catch it when you look inside the answers themselves.
The AEO metrics that matter
AEO measurement comes down to five signals, and citation rate is the one to obsess over. Get these right and you know whether your AEO performance is moving. Chase vanity numbers and you optimize for nothing.
Here is what each signal tells you, and where teams get fooled.
| Metric | What it tells you | Where it misleads |
|---|---|---|
| Visibility / mention rate | How often you appear across your tracked questions | One good day means little. Watch the trend, not the spike |
| Citation rate | Whether your page is the actual source feeding the answer | Being mentioned is not being cited. Citations are where the work starts |
| Share of voice | Your slice of mentions versus competitors | Rising share on ten questions proves nothing. You need a wide question set |
| Sentiment | Whether AI describes you positively or not | An accurate positive mention beats a vague neutral one every time |
| Recommendation rate | Whether AI actively recommends you, not just names you | The metric closest to revenue, and the one most tools bury |
A mention says AI knows you exist. A citation says AI trusts a specific page enough to build its answer on it, and that page is now doing your selling. That is why citation rate, not raw mentions, is the number we report to clients first, and the first thing an ecommerce AI search audit measures.
How to track AI traffic in GA4 for free
You can see AI referral traffic in Google Analytics today, for free, in about fifteen minutes. Google added a native AI Assistant channel in 2026 that recognizes ChatGPT, Gemini, and Claude, but it misses a lot. So the reliable move is to build your own channel with a simple rule.
Here is the setup, start to finish:
- Open GA4 Admin, then go to Channel Groups under Data Display.
- Create a new channel group and add a channel called "AI Search."
- Set the condition to Source matches regex, then paste a pattern covering the AI domains (chatgpt.com, perplexity.ai, gemini.google.com, claude.ai, copilot.microsoft.com).
- Drag "AI Search" above Referral, because GA4 sorts each visit into the first channel it matches.
- Save, open Traffic Acquisition, and add Landing page as a secondary dimension.
That last step is the one that earns its keep. It shows you exactly which pages the answer engines send people to, and which pages they ignore.
For a Shopify store, this changes the conversation fast. If AI sends traffic to a blog post but never to your collection or product pages, that is a content gap, not a tracking gap.
- Two honest caveats. Free ChatGPT users and in app browsers often strip the referrer, so some AI visits still hide inside Direct traffic. And Google AI Overviews traffic still counts as organic, not as a separate channel. So treat the GA4 number as a floor, not the full picture.
One upside makes the setup worth it anyway. AI referred visitors tend to convert well above regular organic in study after study, though the exact multiple swings too much to quote to your CEO. A vape retailer we work with saw their AI sessions convert noticeably higher than search, which is exactly why the channel deserves its own row.
The best AEO tracking tools worth paying for in 2026
GA4 shows you clicks, but it cannot show you the answers where you get named and nobody clicks, which is most of them. That is where a dedicated AEO tracking tool comes in. It runs your questions against the engines on a schedule and records who gets mentioned and cited.
Here is how the main options compare. Confirm current pricing on each vendor's page before you commit, because this market re-prices constantly.
| Tool | Best for | Notes |
|---|---|---|
| Profound | Enterprise depth and large prompt research | Strong citation data and fast detection. Priced for bigger budgets |
| Peec AI | Growing brands that want focused tracking | Clean mention and competitor tracking without enterprise overhead |
| Otterly | Teams starting structured AEO tracking | Straightforward visibility monitoring across the major engines |
| HubSpot AEO | Marketers already inside HubSpot | Tracks a set number of prompts across ChatGPT, Gemini, and Perplexity |
| Ahrefs Brand Radar | SEO teams adding AI to an existing stack | Sits next to the rank tracking you already run |
You do not need one on day one. Start with the free GA4 setup and a short list of prompts you run by hand, then upgrade when the budget is on the line — see what AEO cost looks like by tier. Honestly, the discipline matters more than the tool. A cheap platform checked weekly beats an expensive one you open once a quarter.
How often should you track AEO?
Track AEO weekly, against fifty to a hundred and fifty questions, across more than one engine. Do less than that and your data will lie to you. This is the part almost every guide skips, and it is the part that decides whether your numbers mean anything.
- Weekly, not once. AI answers are non deterministic, so the same question can return two different answers. Running a prompt once and calling yourself visible is a coin flip you happened to win.
- Fifty to a hundred and fifty questions. Ten prompts will not give you a stable signal. Most ecommerce categories need that range to cover how real buyers actually ask.
- More than one engine. Each engine cites differently, so you can win Perplexity and be invisible in AI Overviews at the same time.
From our work: A Finnish sauna store we track was cited steadily in Perplexity while showing up almost nowhere in Google's AI Overviews. Only a multi engine view made that gap obvious, and it changed where we pointed the content next.
Turn your AEO tracking into action
Tracking only pays off when it changes what you publish. The whole point is to spot the gap, then close it. When your citation data shows an answer engine pulling from a competitor's buying guide instead of yours, you know the exact page to rebuild.
- Find the cited source. See which page (yours or a rival's) is feeding the answer.
- Rebuild yours to beat it. Make your page the clearer, better sourced answer to that question.
- Watch the citation shift. Track the same questions weekly and confirm the fix landed.
Fix the page the answer is built on, and you do not just win one query. You start showing up across every related question that page can feed, which is the topical authority for AEO you are tracking for.
Track AEO like you track revenue, not like you track a hunch. The teams winning AI search are not the ones with the fanciest dashboard. They are the ones who check weekly, spot the gap, and go fix the page.
Want this run on your store? Book a free 30 minute walkthrough.