Menchly research note
The recommendation gap: known is not selected
Why awareness can coexist with exclusion from an AI-assisted shortlist - and how leaders can diagnose the difference without claiming control over independent systems.
Executive summary
A brand can be prominent in its category yet absent when an assistant is asked to choose for a particular buyer, mission or constraint. Awareness answers “what exists?” Selection answers “what fits this decision, and why?” The second question demands specific, current and supportable evidence.
A brand can be prominent in its category yet absent when an assistant is asked to choose for a particular buyer, mission or constraint Awareness answers “what exists?” Selection answers “what fits this decision, and why?” The second question demands specific, current and supportable evidence.
Working definitions
Language leaders can use precisely.
“The recommendation gap is the distance between being recognisable as a category participant and being supportably selected for a defined buyer, need and decision context.”
“Selection evidence is current, attributable information that helps a buyer distinguish credible fit - not simply proof that a brand exists or is popular.”
Awareness and selection solve different questions
Awareness signals can establish category membership: a known name, repeated press mentions, a substantial social presence or a long operating history. Those signals may help an assistant recognise an entity. They do not necessarily explain when the entity is suitable, what it is unusually equipped to do, or which constraints should rule it in or out.
Selection is narrower. A recommendation request normally contains - or implies - a buyer role, intended outcome, geography, risk threshold, budget logic, service expectation and comparison set. If public information does not connect the brand to those criteria, an assistant may default to better-documented alternatives, generic popularity signals or cautious category summaries.
Awareness signals can establish category membership: a known name, repeated press mentions, a substantial social presence or a long operating history Those signals may help an assistant recognise an entity They do not necessarily explain when the entity is suitable, what it is unusually equipped to do, or which constraints should rule it in or out.
Diagnose the gap across four executive questions
First, presence: does the brand enter answers for a deliberately selected set of material questions? A single appearance is not a stable ranking, and absence from one response is not proof of broad invisibility Use repeated, documented observation as directional evidence.
Second, fit: when the brand appears, is the stated reason aligned with the clients, missions and markets it can genuinely serve? An irrelevant recommendation can create reputational and operational cost even when it appears positive.
Third, accuracy: are entity details, capabilities, locations, ownership relationships and limitations current? A polished answer built on stale facts is not a successful outcome.
Fourth, authority: can material claims be traced to reliable owned records or credible independent sources? NIST’s AI Risk Management Framework emphasises validity, reliability, transparency and ongoing measurement as characteristics of trustworthy AI risk management For a brand team, the practical implication is to treat model output as evidence to evaluate - not as an unquestionable verdict.
Build reasons for consideration, not claims of superiority
A responsible authority programme begins with bounded recommendation territories: the buyer situations in which the brand has a substantiated right to be considered. Teams can then map each territory to approved facts, explanatory owned content, expert credentials, relevant third-party coverage and clear entity relationships.
The discipline is subtractive as well as additive. Claims that cannot be evidenced should be narrowed or removed. Confidential projects should not be disclosed to fill a content gap. Where proof cannot be made public, the brand can often explain process, standards, governance and anonymised capability boundaries without implying a result it cannot demonstrate.
A responsible authority programme begins with bounded recommendation territories: the buyer situations in which the brand has a substantiated right to be considered Teams can then map each territory to approved facts, explanatory owned content, expert credentials, relevant third-party coverage and clear entity relationships.
Important limitations
Interpret with care.
- AI assistant outputs vary by model, version, retrieval method, location, account context and prompt wording.
- The frameworks in this note support strategic judgement; they do not predict or guarantee recommendation, ranking, revenue or reputation outcomes.
- Examples are illustrative composites, not client work, performance claims or evidence that any named or implied brand has adopted this approach.
Sources
Further reading and primary references.
- NIST, Artificial Intelligence Risk Management Framework (AI RMF 1.0)National Institute of Standards and Technology
- OECD AI PrinciplesOrganisation for Economic Co-operation and Development
- Guidelines for Human-AI InteractionMicrosoft Research / CHI 2019
Private AI visibility assessment
Turn the strategic question into a brand-specific baseline
Establish how assistants interpret your brand today, which questions matter, and where infrastructure can improve speed, learning and compounding results.
Request a private assessment