- An AI Shortlist Scan is a specific, named check: 15 buyer-language prompts, run three times each across ChatGPT, Perplexity, Gemini, and Claude, checking whether a brand or a named competitor gets recommended. $497, one time, results in three days.
- The prompts are buyer-language, not branded. It's the same "would a real prospect actually type this" filter covered in building a fixed prompt set, not a keyword list with a brand name baked in.
- Three runs per prompt is a deliberate, stated scope limit: enough to rule out one lucky or unlucky screenshot, not a substitute for the 7-runs-per-day, multi-week rigor a real ongoing measurement program needs.
- It answers one narrow question: are you currently showing up, or is a named competitor. It doesn't answer why, and it doesn't say what to do about it. That's the Baseline Diagnostic's job.
- The whole point is that it's disclosed: the exact prompt count, run count, and engine list are stated up front, the same methodology-transparency standard this site holds vendors to elsewhere.
- More than one firm now uses "AI Shortlist Scan" as a name, for different offers at different prices and depths. This piece defines the specific, disclosed version covered on this site.
Most of what circulates about "AI visibility checks" is either a single unbranded ChatGPT screenshot passed off as proof, or a vague audit with no stated methodology. This piece defines a specific, named check plainly: what actually happens, why it's shaped the way it is, and exactly where its scope stops.
Is this the same "AI Shortlist Scan" other agencies offer?
No. "AI Shortlist Scan" describes the check, not one company's product, and more than one firm now uses the name for a different offer at a different price and depth. This version is a fast, narrow, $497 tripwire: 15 buyer-language prompts, 3 runs each, across ChatGPT, Perplexity, Gemini, and Claude, with results in 3 days, built to catch a directional pattern before committing to a fuller engagement. A pricier, slower version claiming the same name may cover more ground (competitive framing, the post-click buyer experience, a longer walkthrough) at a correspondingly higher price and turnaround. Check what's actually included and how the methodology is disclosed before assuming two things sharing a name are interchangeable.
Being recommended by name isn't a fixed fact about a brand to begin with. A 14,000-API-call Conductor study found brand overlap between repeated runs of the identical prompt ranging from just 40% at the purchase-decision moment to 63% for comparison queries, with lead-brand retention ranging from 59% to 91% across the same span (see being cited vs. being chosen by AI for the full breakdown). The shortlist a buyer sees can look meaningfully different from the one the same buyer would see five minutes later. That instability is exactly why a one-time check, however it's branded, is worth running before assuming you already know where you stand.
What actually happens during an AI Shortlist Scan?
Fifteen real buying-stage prompts, written in the language an actual prospect would use rather than a query with the brand's own name baked into it, get run three times each across four engines: ChatGPT, Perplexity, Gemini, and Claude. That's 180 total runs (15 × 3 × 4). Each response gets checked for whether the brand appears at all, whether it appears against named competitors, and which sources each engine is citing when it answers. The output is a table showing mention rate and shortlist rate against named competitors, plus a short video walkthrough of what the runs actually found. Turnaround is three days; the fee is $497, paid once.
Why buyer-language prompts instead of the brand's own name?
A prompt like "is [Brand] good" tests something almost nobody asks a chat interface. Real buyers describe a problem or a category, not a company they've never heard of yet. Building a fixed prompt set covers the fuller construction method (IAB's four intent categories, and the filter for trimming a draft list down), and the same discipline applies here at a smaller scale: 15 prompts built from what a category's actual buyers would type, not a keyword list wearing a question mark.
Why three runs per prompt, not more?
A single run of a single prompt is close to a coin flip. How many prompts do you actually need for a trustworthy read? covers the two independent studies behind that claim directly: SparkToro and Gumshoe.ai found under 1-in-100 odds that two runs of an identical prompt return an identical brand list, and the University of St. Gallen's 45-day study found identical prompts run simultaneously overlapping only 32–43%. Running each prompt three times across four engines (180 total runs) is enough to tell a real, repeatable pattern from one lucky or unlucky screenshot, which is what most "I checked and we're there" claims circulating online actually are.
Is three runs the same rigor as a real ongoing tracking program?
No, and this is a stated scope limit, not a gap to paper over. The same piece that establishes the three-run rationale also states the actual floor for an ongoing, trend-level measurement program: at least 7 runs per prompt per day, and a 2–4 week rolling window, per the University of St. Gallen's own convergence analysis. An AI Shortlist Scan is a one-time snapshot built to catch a directional pattern (invisible everywhere, or losing consistently to one named competitor), not a continuously-sampled, statistically rigorous trend line. If what's actually needed is the latter, the arithmetic for that is covered separately, and it's a different kind of engagement than a $497, three-day check.
Where does it sit next to the Baseline Diagnostic and an ongoing program?
The Scan answers one narrow question: are you, or a named competitor, currently showing up for real buying-stage prompts. It doesn't explain why, and it doesn't fix anything. The Baseline Diagnostic goes further: it audits crawler access, entity hygiene, content extractability, and earned-media presence across all four stages to explain the mechanism behind whatever the Scan found, and prices a plan to fix it. An ongoing tracking program, built around the repeated-sampling discipline covered in how many prompts do you actually need, is the version built for a trend rather than a one-time read. The Scan exists as the fastest, cheapest way to find out if there's a problem worth taking further. It isn't a substitute for either of the other two.
Frequently asked questions
What is an AI Shortlist Scan?
A one-time, $497 check on whether ChatGPT, Perplexity, Gemini, and Claude recommend your brand or a named competitor for real buying-stage questions: 15 buyer-language prompts, run three times each across all four engines, with results in three days.
Is this the same "AI Shortlist Scan" other agencies offer?
No. AI Shortlist Scan describes the check, not one company's product, and more than one firm now uses the name for a different offer at a different price and depth. This version is a fast, narrow, $497 tripwire: 15 buyer-language prompts, 3 runs each, across ChatGPT, Perplexity, Gemini, and Claude, with results in 3 days, built to catch a directional pattern before committing to a fuller engagement. A pricier, slower version claiming the same name may cover more ground (competitive framing, the post-click buyer experience, a longer walkthrough) at a correspondingly higher price and turnaround. Check what's actually included and how the methodology is disclosed before assuming two things sharing a name are interchangeable.
Why 15 prompts, three runs each, across four engines specifically?
Fifteen buyer-language prompts is enough to cover a category's real buying-stage questions without diluting the sample across queries nobody actually asks. It's the same filtering discipline covered in building a fixed prompt set. Three runs per prompt isn't a rigorous statistical sample on its own; it's enough to tell a real pattern from one lucky or unlucky screenshot, which is what most "I checked ChatGPT and we're there" claims actually are.
Is a 3-run scan the same rigor as the 7-runs-per-day standard covered elsewhere on this site?
No, and that's a deliberate, stated scope limit, not an oversight. How many prompts do you actually need for a trustworthy read? covers the University of St. Gallen's recommendation of at least 7 runs per prompt per day and a 2-4 week rolling window before trusting a trend. That's the bar for an ongoing measurement program. An AI Shortlist Scan is a one-time snapshot, not a trend. It's built to catch a directional pattern (invisible everywhere, or losing consistently to one named competitor), not to serve as a statistically rigorous, continuously-tracked score.
How is this different from the Baseline Diagnostic or an ongoing tracking program?
Scope and depth, not accuracy. The Scan checks one specific question (are you or a named competitor showing up for real buying-stage prompts) and stops there. The Baseline Diagnostic goes further, auditing crawler access, entity hygiene, extractability, and earned-media presence to explain why the pattern exists and what to do about it. An ongoing tracking program (see how many prompts do you actually need) is the version built for a trend, not a one-time snapshot. The Scan is the fastest, cheapest way to find out if there's a problem worth investigating further.
Sources referenced
- Schulte, Bleeker & Kaufmann, "Don't Measure Once: Measuring Visibility in AI Search (GEO)" (University of St. Gallen)
- SparkToro / Gumshoe.ai AI recommendation consistency study (Jan 2026, ~2,961 prompt runs)
- Conductor: AI Brand Recommendation Study — Why Intent Type Predicts AI Output Consistency (14,000 API calls, 10 industries, 7 intent types, 4 LLMs)
This piece defines the product referenced throughout is your brand on the AI buyer's shortlist before anyone even searches?. For the sampling arithmetic behind an ongoing measurement program (as opposed to this one-time check), see how many prompts do you actually need. To run the actual check, see the AI Shortlist Scan page directly.
