One — entity confusion
The most common cause of AI invisibility is that the AI cannot determine which subject is being described. The business operates under a registered name, a trading name, a Google Business Profile name, and a third-party directory name — and the four are not consistently linked. The system therefore treats the subject as multiple weak entities rather than one strong one.
The fix is entity reconciliation: a canonical entity record signed by the principal, then conformed across every property the system consults.
Two — authority distance
The second-most common cause is that the subject is too far, in the citation graph, from the authorities the AI weights. The business may be excellent operationally and reputationally, but its public footprint does not include credible, recent third-party citations from the sources the system regards as authoritative.
The fix is to close authority distance deliberately: earned editorial coverage in trade press, professional body membership made machine-readable, credentialed authorship architecture across the property.
Three — citation absence
The third cause is the absence of citation-quality content on the subject's own pages. The pages exist, are indexed, and are reasonably authoritative — but the substantive answers are buried in long marketing copy, unattributed, and not quotable. The AI cannot extract a passage to cite, so the page is invisible at the citation moment even where it ranks in classic search.
The fix is editorial: existing pages rewritten so that substantive answers live in short, attributed, evidence-bearing passages.
Four — editorial mismatch
The fourth cause is voice. The business writes about itself in a register that AI systems read as promotional rather than substantive. The content may be technically accurate, but it does not read as if produced by an institution. The AI is, in this respect, unusually sensitive: subjects whose voice oscillates between marketing and substance are weighted lower than subjects whose voice is consistently editorial.
The fix is voice consistency. The subject's content is rewritten in a single, restrained, evidence-led register, and held to that register going forward.
Most subjects who arrive citing AI invisibility carry deficits across two, three, or all four of these causes. The institution's audit identifies which deficit dominates and sequences the work accordingly. The work is rarely glamorous; the lift, when it arrives, is durable.
What to do, in order
- Audit the subject against the eight Discovery Signals to identify which deficits dominate.
- Reconcile the entity across all properties the system consults; this is the prerequisite to every later move.
- Close authority distance through credentialed authorship and earned editorial coverage.
- Rewrite existing pages for citation-quality content, not for additional content volume.
- Hold the voice. Editorial consistency compounds; oscillation does not.
A closing observation
AI invisibility is not, in nine cases out of ten, a mysterious problem. It is the cumulative consequence of four structural deficits, none of them new, none of them solved by more marketing.
The subjects that succeed at becoming visible are the subjects that take the diagnosis seriously and accept that the work is closer to estate maintenance than to a campaign. The lift, when it arrives, is the lift of a property restored.