AI‑Tracker: discover whether AI recommends your business

AI‑Tracker measures whether ChatGPT, Google Gemini and Anthropic Claude independently recommend your business—or select your competitors instead.

Unlike checks that include the company name in every prompt, AI‑Tracker begins with 20 realistic customer questions that do not name the business.

Those questions are tested independently across all three AI assistants, producing 60 measured responses.

The resulting report shows:

• Whether your business was recommended
• How often it appeared first
• Which competitors were selected
• Why those competitors may have been chosen
• How results differed between AI providers
• What your business should improve next

General informational answers are separated from genuine supplier-selection answers, so a business is not marked as having lost when no supplier was recommended.

One-off reports and monthly monitoring are available for UK and US businesses.

getaitracker.com

We don’t guess what AI thinks about your business. We measure it.

AI‑Tracker: discover whether AI recommends your business

This is a sensible angle for UK SMEs, because a lot of “AI visibility” tools quietly bake the brand name into prompts and then claim a win. Starting with unnamed, real customer-style questions is closer to how prospects actually search.

A couple of practical points buyers will ask:

• Do the prompts vary by UK location/intent (eg “near me”, “Manchester”, “emergency”, “B2B vs consumer”)?
• How do you control for model randomness—fixed temperature, repeat runs, and timestamping?
• Can the report separate “no supplier suggested” from “competitor suggested”, and show the exact citations/phrasing used?

If those are covered, it’s a useful benchmark alongside normal SEO and reviews. Link: getaitracker.com

Thank you...you have identified three of the most important methodological issues.

Yes, the questions are generated for the individual business, industry and stated location. They include different commercial intents rather than repeatedly rephrasing one generic query, including local, trust, comparison, value and supplier-selection questions where relevant.

Model variability cannot be eliminated, so AI‑Tracker does not present an individual response as a permanent ranking. Each scan is timestamped, the same 20 questions are tested independently across ChatGPT, Gemini and Claude, and repeated scans provide the more meaningful comparison over time.

The report also distinguishes between:

• The business being recommended
• A competitor being recommended
• Suppliers being compared
• A general informational response in which no supplier was selected

That last distinction is crucial, a business is not recorded as having “lost” when the AI did not recommend anyone.

The full report retains the provider-by-provider evidence and wording used to support the findings, rather than presenting an unexplained headline score.

I appreciate the thoughtful questions.