How We Measure AI Visibility

Here is how our checks work, what we record, and what the results can tell you about your firm.

What we test

We send three questions about a legal service and location to the providers’ developer tools. A model is the software that generates the answer; an API is how our tool communicates with it. These checks are not a live recording of the consumer ChatGPT, Claude, or Gemini apps. Those apps can use different models, search tools, context, and personalization.

What the counts mean

A count shows how many tested answers recommended a business. It is not a search position, market share, or estimate of lost clients. A provider that fails is excluded and identified as not measured. An absent recommendation means only that we did not observe it in those answers.

How to read the report

New reports save the questions, original answers, model names, and whether each check completed successfully. Older reports explain when the original answers were not saved. We create a new record for a follow-up rather than rewriting the previous results.

Business discovery

Google Places supplies the initial business sample. Places result order is not organic Google search ranking. Firms without websites and duplicate domains can be excluded, so this is a sample rather than an exhaustive list of firms in the city.

What happens next

We review the evidence and agree on specific improvements to your online information. As part of monthly support, we repeat the agreed questions and review how the answers change. This helps us decide which improvements to work on next. AI answers vary; no individual result is a guarantee of future recommendations.

Curious where your business stands? See a sample report or get in touch to run yours.