Decision evidence
Mentioned is not chosen
Preference separates raw visibility from recommendation risk by asking providers to compare the client against named competitors inside the buyer context.
Find the competitor AI would choose before the client sees the shortlist.
AI visibility is not the same as Preference. A client can be mentioned and still lose the recommendation when ChatGPT, Claude, Gemini, Perplexity, Grok, Mistral, or DeepSeek is asked which vendor to choose. VectorGap turns those direct-matchup answers into category scores, win/loss reasons, evidence gaps, remediation missions, retests, and a report your agency can use to explain whether the competitor won because its public proof was easier to retrieve, cite, or trust.
- Direct brand-versus-competitor matchups instead of mention counting.
- Category evidence for price/value, capabilities, innovation, implementation ease, trust, market fit, support, and overall recommendation.
- Provider-by-provider win/loss reasons with model-memory, citation, weak-source, and unsupported-claim flags.
- Mission Control handoff, comparable retests, and client-ready reports for agency delivery.
40
preference prompts
8
buying categories
7
standard providers
Positioning gap
Preference needs market proof
Models often choose the brand whose evidence is clearer, fresher, and easier to cite. VectorGap shows whether the loss is a true preference issue or a public-proof gap the agency can repair.
Agency output
Loss reason → mission → retest
Every category loss can become a concrete fix: strengthen proof, repair source evidence, clarify entity facts, add comparison-safe claims, then rerun the same target.
A brand can appear in the answer and still lose the deal
Most AI visibility tools stop at mentions. That is too weak for client strategy. Agency buyers need to know whether AI would choose the client when a buyer asks for the best option, the safest vendor, the strongest value, or the easiest implementation.
Direct recommendation risk
VectorGap tests forced-choice and shortlist-style prompts so the agency sees whether the client wins, loses, or gets buried behind a named competitor.
- Primary rank
- Winning brand
- Direct win rate
Provider variance
ChatGPT may choose one competitor while Claude or Perplexity prefers another. That disagreement reveals where the public proof is strong, fragile, or provider-specific.
- Provider-level ranks
- Answer excerpts
- Consistency score
Buyer-context targeting
Preference is evaluated against the market, language, industry, persona, and competitor set the client actually sells against, not a generic global category.
- Target market
- Persona intent
- Competitor set
Eight buying categories show why AI chooses the competitor
A single preference score is not enough. VectorGap breaks the matchup into the commercial reasons buyers care about, then links weak categories to the evidence the agency can improve.
Price and value
See whether AI frames the client as worth the spend, too expensive, unclear, or less proven than alternatives.
Capabilities and innovation
Find where the client’s features, roadmap, differentiation, or future-readiness are under-explained in public evidence.
Ease, trust, market fit, and support
Expose the categories that often decide B2B recommendations: implementation confidence, proof of credibility, target-market fit, customer success, and the final overall recommendation.
The useful part is the reason behind the win or loss
Preference is valuable because it preserves the answer evidence. The agency can inspect the provider, category, rank, rationale, unsupported claims, citations, and decisive reasons, then compare the answer against the public proof that made the competitor look safer.
Decisive reasons
Show the client why a competitor looked safer: clearer positioning, stronger proof pages, third-party authority, category-specific evidence, or better entity consistency.
Unsupported-claim flags
Separate real public proof from model guesswork so the agency does not build a strategy on hallucinated competitor strengths or invented client weaknesses.
Source and citation gaps
Identify whether the provider relied on stale AI memory, weak public sources, stale pages, or missing client facts that need to be made extractable.
Preference loss → report proof → Agency OS
Preference diagnostics only become buyer value when the loss reason turns into report evidence, checkout-risk reduction, and recurring delivery capacity. Route the buyer from the matchup proof into pricing proof-before-payment, then into Agency OS when they are ready to run the same workflow across a client portfolio.
Inspect the matchup report
Show how direct competitor losses become provider evidence, category scores, source gaps, Mission Control actions, and a client-ready retest narrative.
Open pageReduce payment risk
Pricing proof-before-payment lets the agency inspect the GBI, Preference evidence, Mission Control actions, and report shape before it commits to the operating plan.
Open pageBuy the operating capacity
Agency OS turns preference diagnostics into recurring matchup audits, remediation missions, same-target retests, exports, and client-ready reports.
Open pageWhich evidence does AI need before it can reuse the page?
This table turns the page into a structured extraction target: the buyer question, the evidence an AI system can read, and the action an agency can sell or execute next.
| Buyer question | What AI can extract | Agency action |
|---|---|---|
| Is the client mentioned or actually chosen? | The page separates raw visibility from competitor preference by showing direct matchup prompts, category scoring, provider rationales, and direct win/loss reasons. | Run Preference when the client needs to know why AI recommends a competitor, not just whether the brand appears. |
| Which proof gap caused the loss? | Price/value, capabilities, trust, market fit, implementation ease, support, innovation, and overall recommendation categories are tied to answer evidence and source gaps. | Turn the losing category into a remediation sprint: comparison-safe proof, source strengthening, schema/entity cleanup, and category-specific evidence. |
| How does the agency prove movement? | Same-target retests preserve market, language, persona, providers, categories, and competitors so the next client-ready report can compare preference movement. | Rerun the exact matchup after fixes and report whether the recommendation reason, rank, and evidence support changed. |
Questions agencies ask before turning AI visibility into client work
How is Preference different from Perception?
Perception shows how providers describe, cite, and understand a brand. Preference asks a harder commercial question: when the provider compares the client against named competitors, which brand does it recommend and why?
Can an agency use this without promising placement?
Yes. The safe offer is evidence, not guarantees: direct matchups, provider rationales, source gaps, missions, and same-target retests. Agencies should sell the diagnostic and remediation workflow, not a promise that models will choose the client.
What should the agency fix after a preference loss?
Start with the loss reason. Typical fixes include clearer category proof, comparison-safe positioning, third-party source strengthening, FAQ/schema/entity cleanup, and better client-ready evidence for the category where the competitor won.