AI systems either include your business in answers or exclude it entirely. Most businesses do not know which. The gap between invisible and named is not a content problem or an SEO problem. It is a structural problem governed by a dependency chain that most businesses have never mapped.
If you are not named, you are not contacted.
- Retrieval — can AI systems find you?
- Extraction — can they cite your content?
- Selection — will they name you as a provider?
- Floor-by-floor diagnosis across ChatGPT, Perplexity, Google AI Overviews, and Copilot
- Competitor comparison — who is cited where you are not
- Priority actions ranked by impact
- CITATE structure — content rewritten for extraction and attribution
- Entity clarity — your business defined consistently across systems
- Page architecture — sections designed to be cited, not just read
- AI systems can retrieve you reliably
- Content sections pass citation criteria C1–C6
- Your entity is defined, not assumed
- Third-party corroboration — independent references to your business
- Entity reinforcement — consistency across platforms and sources
- Ongoing content aligned to how AI decomposes queries
- Inclusion in AI-generated shortlists
- Named recommendation, not anonymous source use
- Compounding advantage as citations become training signals
Why this sequence is not optional
AI visibility is governed by dependency, not optimisation. Businesses do not fail gradually — they fail at the first point where the answer becomes “no.” The correct strategy is not to improve everything. It is to identify the first failure point and resolve it completely before moving to the next. These layers are dependent, not linear. Improving higher layers without resolving lower constraints does not produce results. Improving the wrong layer does not improve results — it hides the real failure.
Can AI systems retrieve your content?
Retrieval depends on technical access, Bing indexation, and crawlability. Without it, everything downstream is irrelevant. If the answer is no, you are invisible. Nothing else matters until this is resolved.
Can they extract and cite your content?
Extraction depends on content structure, standalone answers, attributed statistics, and entity clarity. Retrievable content that cannot be extracted is used anonymously or not at all. If the answer is no, you are retrieved but unnamed. Your content informs the answer. Your business is not credited.
Will they name you as a recommended provider?
Selection depends on third-party corroboration, cross-platform entity consistency, and editorial trust signals. Content quality alone does not govern this layer. If the answer is no, you are cited but not selected. Competitors with equivalent content and stronger entity signals take the named position.
Limits of this model
This model defines where failures occur and in what sequence to resolve them. It does not guarantee outcomes. AI selection is also influenced by signals outside content and structure — domain authority, brand familiarity in training data, competitor presence, and platform-specific behaviour all interact with the work done across these three stages.
The CITATE framework improves extractability and attribution confidence. It does not guarantee selection. The three-floor model diagnoses the failure point. It does not predict how long resolution takes. The Selection Layer describes how AI systems choose. It does not control platform-specific weighting.
How this compares
Most businesses approach AI visibility through one of three incomplete strategies. Traditional SEO produces rankings but no AI presence. A content programme without CITATE structure produces volume but not citation. An audit without implementation identifies gaps without closing them. The Diagnose → Fix → Win sequence is the only path that produces named recommendation — the commercially decisive outcome — and it is the only path where each stage compounds the one before it.
The key distinction is between being used and being named. A high-ranking site with strong traffic can still produce zero attribution in AI answers — its content informs responses without its business being identified as the source. That is the gap this model closes.
