You cannot see a recommendation you never received. Traffic reports only show the visits that happened, not the shortlists you were left off. So measuring answer engine optimisation starts with the answers themselves, and only then connects them to traffic and leads.
Step one: build the question set
Everything rests on a fixed set of questions, asked the same way each time. A good set:
- Uses the words buyers use, taken from sales calls, enquiries and customer interviews.
- Covers the stages of a decision: understanding the problem, comparing options, choosing a provider.
- Includes the specifics that change the answer, such as location, sector, company size or budget.
- Leaves out your brand name for the core measure. Branded questions ("what is Cited?") are tracked separately.
Twenty to fifty questions is a workable size for most companies. Keep the set stable so changes over time mean something, and add new questions as a separate group rather than swapping old ones out.
Step two: run it across assistants, repeatedly
Ask every question in each assistant that matters to your buyers: usually ChatGPT, Google AI Overviews and AI Mode, Gemini, Perplexity, Claude and Copilot. Answers vary from one run to the next, so a single check is an anecdote. Run the set regularly, ideally daily or weekly with a tracking tool, and look at the pattern over weeks rather than any one answer.
The metrics
Mention rate
The share of answers that name your company. This is the headline measure and the one to track per assistant and per group of questions.
Position in the answer
Being named first with a reason is worth more than appearing fifth in a list. Track where you appear and whether the answer recommends you or merely lists you.
Description accuracy
Read what the answer says about you. Is the service right? The location? The type of client? Wrong descriptions send the wrong buyers or put off the right ones, and they point to inconsistent facts across your sources.
Share of voice
How often you are named compared with the competitors who appear in the same answers. This shows whether you are gaining ground or whether the whole category is simply being mentioned more.
Cited sources
Which websites the assistants cite when they answer your questions. This tells you where to be present, and whether your own pages are being used.
Connecting visibility to traffic and leads
- Referral traffic. In Google Analytics 4, visits from chatgpt.com, perplexity.ai, claude.ai, gemini.google.com and copilot.microsoft.com appear as referrals. ChatGPT also adds utm_source=chatgpt.com to many links. Create a channel group for AI assistants so the trend is visible.
- Google's AI features. Clicks from AI Overviews and AI Mode are counted within Search Console's normal web performance data and are not reported separately.
- Ask your leads. Add "How did you hear about us?" to enquiry forms and include AI assistants as an option, with a free-text field. Many buyers who found you through an assistant arrive by typing your name into Google, which analytics records as branded search.
- Branded search trends. Rising branded search in Search Console alongside rising AI mentions is a reasonable sign that answers are sending people your way, though it is not proof on its own.
Reporting that helps decisions
A useful monthly report answers four questions: where are we named now, what changed since last time, which sources drove the change, and what we will work on next. Long lists of vanity numbers do not help anyone decide anything.
Mistakes to avoid
- Judging progress from one person's ChatGPT account.
- Changing the question set every month, which makes trends meaningless.
- Treating AI referral traffic as the whole picture. Many AI-influenced buyers never click a link in the answer.
- Reporting mentions without reading how you are described.
Frequently Asked Questions
What is the most important AEO metric?
Mention rate on a fixed set of unbranded buyer questions, tracked per AI assistant. It shows directly whether you are part of the answers that shape shortlists.
How do I track traffic from ChatGPT?
In Google Analytics 4, look for referrals from chatgpt.com and links tagged utm_source=chatgpt.com. Group these with other AI assistants in a custom channel group.
Which tools track AI visibility?
Dedicated tools such as Profound, Peec AI, Otterly and Searchable run question sets across AI assistants automatically, and large SEO platforms including Semrush and Ahrefs have added AI visibility features. A spreadsheet and manual checks work for a first baseline.
How often should AEO be measured?
Track automatically at least weekly, because answers vary from run to run, and review the trend monthly.