AEO measurement should separate what you observed in an answer, what happened on your website, and what happened in the business. A citation can be useful evidence of visibility, but it is not the same as a qualified lead or a sale.
Build a repeatable observation process, use available native reports, and connect website activity to your CRM. This produces a more useful picture than a single score labeled “AI visibility.”
Define the questions that matter commercially
Create a focused set of questions that represents the decisions your customers make. Include service discovery, provider comparison, pricing, suitability, and process.
Separate branded questions from nonbranded questions. Asking an answer system about your company by name tests a different situation from asking it to recommend an appropriate type of provider.
Document why each question belongs in the sample. A question should connect to an offer or an important customer concern.
This sample is a measurement panel, not the entire universe of customer searches.
Record enough context to repeat the observation
For each test, record the platform, date, exact question, relevant location settings, and whether search or another mode was used. Retain the response and cited URLs when available.
Use fresh sessions where practical and avoid repeatedly prompting the system until it produces the result you wanted.
Repeat observations over time and note that answers may vary even when the question stays the same. A change in the sample does not establish a permanent ranking.
The purpose is to identify patterns and information gaps, not manufacture a favorable screenshot.
Distinguish different kinds of visibility
| Observation | What it tells you | What it cannot establish alone |
|---|---|---|
| Brand mentioned accurately | The response included a correct reference | That the user clicked or contacted you |
| Website cited | A page was used as a linked source in that response | The total reach of that citation |
| Brand described incorrectly | There is an information issue to investigate | Which single source caused the error |
| Referral session recorded | A visit arrived with identifiable referral information | Every visit influenced by AI |
| Qualified inquiry recorded | A relevant prospect contacted the business | That one visibility change caused the inquiry |
Keep these measures separate in reports. Combining them into one total hides the difference between exposure and commercial response.
Use native platform reporting where available
Google provides a Generative AI performance report for impressions in supported search features. Review its definitions and availability for your property; use analytics and business records separately to assess visits and inquiries.
Use platform reporting to complement your question sample. The sample shows what you observed under controlled conditions; native reporting can provide a different view of real activity within the platform's coverage.
Neither source should be assumed to represent every AI product or every customer journey.
Maintain notes on reporting changes so a new metric or feature does not look like an unexplained performance jump.
Connect referrals to the quality of the visit
In analytics, inspect identifiable referrals from relevant platforms and the landing pages they reach. Then examine whether those visits produce meaningful actions.
Be careful with attribution. Some visits may lose referral information, and people may discover a business in an answer before returning through search or a direct visit.
Use a customer-source question in the inquiry process when appropriate, while recognizing that self-reported answers also have limitations.
The best picture often combines analytics, CRM records, and what prospects say during conversations.
Evaluate changes with a clear hypothesis
Suppose your company is repeatedly described as offering a service you discontinued. You update the website and relevant profiles, then monitor whether the description changes.
That is a concrete information-quality hypothesis. It is more actionable than a general goal to “increase AI authority.”
For content changes, record the page, question group, date, and expected effect. Avoid changing every page at once if you need to understand which improvements were useful.
Our GEO audit checklist helps organize the findings that can lead to these tests.
Report decisions as well as numbers
A useful monthly report explains what was observed, what appears commercially relevant, what remains uncertain, and which action follows.
If visibility increases but inquiries do not, inspect the audience, cited page, and next step. If inquiries improve, review their quality before scaling the work.
Sector 13 helps businesses connect AEO reporting to broader search and lead-generation decisions. Email Josh with your current reporting setup. For the strategic context, read our AEO versus SEO guide.