An illustrative example: measuring AI visibility for an online and instore camera retailer
Imagine an online and instore camera retailer that sells cameras, lenses and accessories, with knowledgeable staff who help customers choose equipment.
The retailer wants to understand whether it appears when people ask AI assistants where to buy, which alternatives are recommended, and whether answers accurately describe its online and store services.
Start with questions customers might ask
For this example, the assessment includes 50 discovery questions that do not name the retailer, covering different buying needs:
- “Where can I buy my first camera and get help choosing a suitable lens?”
- “Which camera retailers sell used equipment with a clearly explained warranty?”
- “Where can I try a mirrorless camera before buying?”
- “Which camera shops offer trade-in towards a new camera?”
Questions about visiting a shop specify a relevant location. Online shopping questions use the retailer’s delivery market. This keeps the assessment relevant to customers the business can actually serve.
A separate set of branded questions examines how the retailer is described when named directly, including its store locations, services and buying policies.
Read the results with a clear denominator
Suppose the assessment returns 50 complete, assessable discovery answers:
| Measure | Illustrative result |
|---|---|
| Answers mentioning the retailer | 10 of 50 |
| Mention rate | 20% |
| Answers explicitly citing the retailer’s website | 6 of 50 |
| Website citation rate | 12% |
The 20% mention rate means the retailer appeared in 10 of the 50 tested answers. It does not mean it appeared in 20% of all camera shopping conversations.
A mention also needs context. An answer might recommend the retailer, include it as an alternative, or simply refer to it. Those are different findings.
An explicit citation points readers to a supporting source. A page recorded as retrieved during research is kept separate from a source explicitly cited in the answer.
If an answer is missing or cannot be assessed, that gap is reported rather than counted as an absence.
Investigate a specific gap
Consider this illustrative discovery question:
- “Where can I try a mirrorless camera and get help choosing lenses in [town]?”
Suppose the answer recommends other shops and does not mention the retailer.
That establishes an absence for this question in this test. The next step is to investigate the evidence: which shops appeared, which sources were cited, and how clearly the retailer explains its own relevant services.
In this fictional scenario, the retailer offers camera demonstrations and lens advice, but its store page only lists an address, opening hours and contact details.
That creates a practical content opportunity. It does not, by itself, prove why the retailer was omitted.
Keep factual follow-ups separate
A separate assisted follow-up could supply verified information about the retailer’s demonstrations and advice service, then ask whether it meets the shopper’s requirements.
If the response then includes the retailer, that shows how the answer responds to the supplied facts within that conversation.
The original discovery result remains an absence. The assisted response is reported separately and does not establish permanent learning or improved visibility in future independent answers.
Turn the finding into a brief
For this scenario, a resulting brief could be:
| Item | Illustrative brief |
|---|---|
| Business need | Help shoppers understand where they can try equipment and receive advice |
| Relevant page | The retailer’s local store page |
| Proposed change | Explain available demonstrations, lens advice, booking requirements and any limitations |
| Owner | Retail marketing manager |
| Approval | Store manager confirms the service details before publication |
| Completion check | The published page accurately explains the service and how to access it |
| Later measurement | Repeat the relevant discovery questions and inspect mentions, recommendations and citations |
The content should help customers make a decision and accurately represent what the store offers.
Check visibility and commercial outcomes separately
Later assessments should retain the same questions, geography and testing method where possible, with any changes documented. Repeated observations help distinguish a continuing pattern from a single answer.
Where suitable analytics are connected, the retailer can also examine identifiable AI referral visits, online purchases and recorded enquiries. Store visits and instore sales require suitable measurement of their own.
An increase in mentions would be a visibility finding. An increase in attributed purchases would be a separate commercial finding. Neither alone proves that a particular page change caused additional sales.
See how this works for your own brand.