AI platforms don't search. They decide. When a prospect asks ChatGPT or Perplexity which private aviation management company to trust, which wealth advisory firm handles complex multi-generational planning, or which concierge medicine practice serves executives in their market — they receive a recommendation, not a list. The firms in that answer did not get there because they are the best. They got there because they gave the machine enough clear, credible, consistent signal to construct a confident answer.
Share of Model is the percentage of relevant AI-generated answers that include your company. Most premium service firms have never measured it. The ones that have are rarely satisfied with what they find.
The gap between what a firm has built and what AI can confirm is always structural. It is always traceable to the same seven failure points. I call them the Seven Pillars of AI Signal.
The Seven Pillars
AI platforms must resolve your company as a single, verified, current business entity. Inconsistent NAP data, outdated directory listings, acquired domains with prior history, and absence from foundational knowledge sources — Wikipedia, Wikidata, Google Knowledge Graph — cause AI to improvise. What it improvises is frequently wrong. The goal is a Golden Record: one dominant, accurate source of truth for your company's identity in the AI knowledge layer.
AI grounds its recommendations in sources it trusts: industry publications, authoritative directories, editorial coverage. The companies with the strongest citation footprints across those sources are the companies AI recommends most confidently. An 18-year operator with top-tier safety certification can be invisible in AI answers because it has never appeared in the publications AI uses as citation references for its category. Editorial and Reference citations carry the most weight. Most premium service firms rely almost entirely on their own website — and that is never enough.
AI crawlers access websites differently than Google does. A JavaScript-rendered site that loads seamlessly in a browser may return a blank page to an AI crawler — and when that happens, no content is extracted regardless of quality. Additional failure points: HTTP rather than HTTPS, misconfigured robots.txt, absent llms.txt, PageSpeed below 40. A technically inaccessible site is whispering in a soundproof room.
Schema markup is a machine-readable identity layer — a structured fact sheet telling AI what your company does, who leads it, where it operates, and what its credentials are. Without it, AI guesses from fragments. Organization schema, Person schema for leadership, Service schema, and LocalBusiness schema together give AI a verified reference it can cite with confidence. Schema cannot be audited until Pillar 3 is resolved — a site that cannot be crawled cannot have its schema tested.
AI answers questions by extracting declarative sentences from content it can access. Brand language — "we deliver exceptional experiences for discerning clients" — gives AI nothing to extract. Declarative language — "Silver Air provides aircraft management services for Gulfstream G650 and Bombardier Global 6500 owners operating from Van Nuys and Santa Barbara airports" — gives AI an answer it can surface, attribute, and recommend. Most premium service firms have invested heavily in content that communicates perfectly to a human reader and communicates almost nothing to the machine that precedes that reader.
Review volume, recency, sentiment, and platform distribution signal to AI that a company is actively operating and trusted by clients. An unclaimed Google Business Profile with no reviews tells AI this company is unverified. In categories where the purchase decision is high-trust and high-value — which describes every vertical LockePoint serves — an unverified company does not get recommended. Hallucination monitoring is equally essential: AI misrepresentation of a firm's services, leadership, or credentials must be identified and corrected before it reaches a prospect.
Competitive AI Presence is the aggregate outcome of all six preceding pillars — and the metric that makes the gap undeniable. When a category query returns three competitors and excludes a firm with stronger credentials than two of the three, the conversation about AI visibility stops being abstract. Every month the gap exists, the companies appearing in AI answers receive inbound interest that belongs to you. The query ran. The recommendation was made. Your competitor got the call.
"A company can rank on page one of Google and be completely absent from AI-generated answers in the same category. SEO and AEO are not the same discipline."
How Each Pillar Is Scored
In every LockePoint AI Signal Audit, each pillar is scored Red, Yellow, or Green. Red means a material gap — an active signal working against the company. Yellow means partial or inconsistent signal — the firm appears inconsistently, and competitors with stronger signal are winning recommendations it should be winning. Green means solid foundation — this pillar is not currently costing Share of Model.
In the firms I have audited, it is rare to find more than two Green pillars at the start of an engagement. Three to four Red pillars is the most common profile — and it produces near-complete AI invisibility on category queries, regardless of how strong the underlying business is.
Where to Start
Run your most important category query on Perplexity right now. Not your brand name — the query your best prospects use when they don't know who to call yet. If your company is not in the answer, you now know the problem exists. The Seven Pillars tell you exactly where to look for the cause.
A 20-minute AI Visibility Snapshot shows you exactly what AI platforms say about your company, what they say about your competitors, and which pillars are costing you the most.
Request Your Snapshot