FIND. TRUST. VALIDATE.
Behind the five pillars sit three blunt questions a model resolves before it names your brand: can I find you, should I trust you, and does the rest of the internet agree? This is the scoreboard we run engagements against.
- Find = Access + Identity + Extractability. Can the model retrieve and read you at all?
- Trust = Authority. Does your content carry the data, citations, and credentials a model leans on?
- Validate = Footprint. Does independent, third-party corroboration back you up - or is it all self-reference?
- Validate is weighted highest, because consensus, not owned media, durably moves visibility.
Can the model find you?
Find is table stakes: if the model can't retrieve, recognize, and read you, the other questions never get asked. It bundles three of the five pillars - Access (it can crawl you), Identity (it knows which entity you are), and Extractability (it can lift a clean answer from your page).
Most brands assume they pass Find and fail it quietly - blocked bots, a JavaScript-only site, or an entity the model confuses with a similarly-named company. The fixes are largely technical and fast, which is why Find is where we start. The mechanics live in the 5 pillars.
Should the model trust you?
Trust is whether your content looks credible enough to quote. It maps to the Authority pillar: original data, hard numbers, named sources, and real credentials. A page with zero data and an anonymous byline is reachable and readable, and still gets passed over for a competitor who backs every claim.
Trust is earned on your own pages but verified against the web - which is where the third question comes in. What specifically lifts trust is covered in what earns an AI citation.
Does the internet validate you?
Validate is the hardest and most decisive question: does the rest of the web agree? It maps to Footprint. A brand can ace Find and Trust on its own site and still lose, because owned media makes you retrievable while consensus makes you recommended.
This is the question most programs underinvest in, and it is why we weight it highest. A mention backed only by your own domain is self-reference; a mention corroborated across forums, reviews, and editorial is consensus. The trap is detailed in why owned media plateaus.
How do the three roll into one number?
Find, Trust, and Validate roll up into a headline AI Presence Score that weights Validate highest. Each AI answer is scored on presence, sentiment, position, and consensus, then aggregated per provider and per prompt category so you can see exactly where you win and where you leak.
The full scoring method - and why the weights are what they are - is in how we measure.