What an AI Visibility Review finds, shown on a real advisory firm
- Client
- A financial advisory firm (name withheld)
- Industry
- Financial advisors
- Scope
- Found, Cited, Answered: search, maps, AI engines
- Delivered
- September 2026
- Status
- Reviews run on ten businesses so far
An AI Visibility Review answers three questions about a business: is it found when people search, is it cited when they ask ChatGPT or Gemini, and is it answered when a client actually reaches out. Each answer is measured, not guessed, and every finding comes with the evidence behind it. This is a real review of a financial advisory firm, run in September 2026 and shown here without its name.
What this review found
| Question | What was measured | Result |
|---|---|---|
| Found | 12 target keywords in organic results and the map pack, plus 49 map grid points on 4 local terms | Not placed on any keyword or grid point |
| Cited | 24 prompt and engine pairs across ChatGPT and Gemini | Never named, in answers carrying 134 competitor mentions |
| Answered | A timed test enquiry by form, phone and message | Not run: it needs the client's consent |
The review turned that into 29 findings, six of them urgent, and a 17-item plan split into quick wins, strategic moves and long plays. The most likely cause, which the review labels as an inference from the evidence rather than a measured cause, was not the website at all. Google's profile for the firm was still under a retired name, with no reviews, no hours and no website link, and one AI engine reads exactly that index.
Five reviews side by side
Every review so far has been different, which is the point of measuring instead of guessing. Three businesses were never named in any AI answer we checked. Two were named often but rarely linked, so the AI talked about them without sending anyone to their site. Only Insurance Pro Shop, a client, is named here; the others are shown by industry.
| Business | Found on Google | Cited by AI | Findings |
|---|---|---|---|
| Independent insurance agency, Southern California | In the map pack on 9 of 15 target terms, #1 on three | Named in 37.5% of answers and the highest share of voice in its category, but linked in only 5.9% | 37 |
| Insurance Pro Shop, sales training for agents | Ranked on 5 of 15 target terms, best at #6 | 50.0% of mentions among five tracked businesses; ChatGPT linked it on 3 of 7 searched questions | 29 |
| Financial advisory firm, Texas (above) | Not placed on any of 12 keywords or 49 map points | Never named in 24 prompt and engine pairs | 29 |
| Regional insurance agency | Absent from all 15 target searches and all 81 map points | Not in any of 5 AI Mode answers or 10 AI Overviews checked | 22 |
| Coach for financial advisors | No non-brand rankings in the top 10 for 15 target terms | Never named in 48 prompt and engine checks across ChatGPT, Gemini, Claude and Meta AI | 19 |
The coach is the clearest example of why both halves matter: hundreds of other websites link to him, which is real authority, and yet no search result or AI answer surfaced him for the questions his clients ask. The insurance agency in California is the opposite: the AI knows it well, and almost never sends anyone to it.
How a review is built
- A keyword universe for the business, with volume, trend, difficulty and intent for every term, so the review measures the searches that actually matter.
- Search and map pack positions for each target keyword, and a map grid showing where in the service area the business appears and where it vanishes.
- Prompt testing across the major AI engines, run in a real browser in more than one pass, recording who gets named and which sources get cited.
- Technical, content, schema, backlink and reputation layers, including Google Business Profile and knowledge panel checks.
- With consent, a timed mystery-shop of the business's own form, phone and messages, because being named only matters if someone answers.
- A dashboard, a written report, a plain-English guide to reading it, and a prioritized action plan.
Why we check our own work twice
Before this review shipped, a second AI model from a different company read the whole report looking for errors. It found 19, including a citation rate the report elsewhere said could not be computed and competitor names that had been counted as mentions because they appeared in our own prompts. All 19 were fixed. Our own checks then caught two results shown as not measured that had in fact been measured at zero. A review that overstates a problem is as useless as one that misses it.
Want to know where you stand?
Start with a free market insight brief, or book a 30-minute working session to talk through a full review.
Get your free market insight brief See more workThe firm's name and identifying details are withheld. Results are specific to one business and are not a prediction for any other.