Person Claim Frame Prove in The Kalicube Process
Published: 14 March 2026 Author: Jason Barnard, CEO of Kalicube® Status: Original concept, first publication
A well-written biography is one of the most compelling things a human can read about another person. The arc of insight, the moments of vindication, the slow accumulation of a distinctive point of view, and a human reader follows that narrative and emerges with a feeling: I understand this person, I trust this person, this person is the real thing. The biography did its job.
The AI read the same piece and extracted almost nothing.
Not because the writing was bad. Because AI doesn’t follow narratives. It counts corroborated statements, and a biography built on implication and arc rather than explicit, independently-repeated factual claims scores close to zero in the entity model, regardless of how well it was written or how true the story is. The human found it convincing. The AI found it uninformative. Both were right.
The human and AI readers are asking different questions
A human reading a person’s credentials is pattern-matching for trustworthiness. They want to know how someone arrived at their position: the thinking, the setbacks, the moments where the insight proved out. They read between lines. They notice whether the story holds together under pressure. A beautifully constructed arc from early observation to vindicated insight is, for a human reader, a powerful proof of genuine expertise.
AI doesn’t read between lines. It aggregates explicit statements, weights them by how independently they’re sourced, and builds a model of the entity from what is directly, repeatedly, unambiguously stated across domains it has no reason to suspect of coordination. The journey doesn’t register. The implication doesn’t register. What registers is: this person’s name, this area of authority, this specific contribution, stated explicitly, appearing on independent sources that had no editorial motive to say the same thing.
Same subject, same underlying truth, different requirements for how that truth needs to be expressed.
Not all person proof carries equal weight
The proof types available for a person’s entity model are not equivalent, and the AI’s weighting diverges sharply from the human’s at several points.
| Priority | Proof type | Human weight | AI weight | Notes |
|---|---|---|---|---|
| 1 | Independent measurement naming the person explicitly | High | Very high | Study, ranking, index - person is subject, not author |
| 2 | Editorial selection by credible publication | High | High | Person chosen, not self-submitted |
| 3 | Peer citation - person’s ideas used by others unprompted | Medium-high | Very high | Explicit attribution language is critical for AI |
| 4 | Self-authored content on third-party platforms | Medium | Medium-low | Platform adds thin endorsement; person is still primary source |
| 5 | Self-authored content on owned domain | Medium-low | Low | Useful for humans already evaluating; AI discounts heavily |
| 6 | Testimonials about the person | Low | Very low | Selection bias obvious to both human and AI |
Two divergences are worth examining closely.
Peer citation sits at medium-high for humans but very high for AI, because when someone else uses your idea and names you as the origin, unprompted, the AI reads that as independent validation of the attribution. No coordination required, no relationship assumed: someone found the idea, used it, credited it. The explicit attribution language is what makes this machine-parseable. “Building on Jason Barnard’s concept of the Entity Home” is a sentence that carries enormous AI weight precisely because it appears on a domain that has no commercial reason to say it.
Self-authored content on your owned domain sits at medium-low for humans and low for AI, a result that consistently surprises people. It’s where you have most control, so it feels like the natural place to build your case. But AI classifies it as first-party content from the entity being evaluated. You are writing about yourself, on a domain you own, in language you chose. The discount is automatic and significant.
Most people don’t have an Entity Home, and that is a structural crisis
Before any of this matters, there’s a prior problem almost nobody talks about. The entire person proof model requires AI to have a resolvable entity for the person: a stable, unambiguous node in its model that it can attach corroborating signals to. Without that node, the signals are orphaned data, floating in the index without connecting to anything.
The Entity Home is what creates the node. A single URL, structured to establish identity, area of expertise, and the key attributions that define the person’s intellectual contribution, in a format the training pipeline can extract and anchor to. Not a LinkedIn profile. Not a company bio page. Not a conference speaker archive scattered across fifteen domains. A personal website, functioning as the authoritative first-party source the AI uses to reconcile everything else it finds.
Most people don’t have one. They built their firm’s digital presence, maintained their LinkedIn, appeared in articles and on podcasts, and never built the home that connects it all. So AI finds fragments (mentions, profiles, partial attributions) and can’t resolve them into a confident entity model. The confidence stays low. The amplification doesn’t fire.
There’s a compounding problem: most names are ambiguous. There are multiple people with the same name, multiple experts in adjacent niches whose signals blur at the edges, and a single person who is genuinely multifaceted: a practitioner in a technical field who is also a speaker, a writer, a founder, and something else entirely that doesn’t fit the category AI is trying to resolve. AI struggles to hold all of those facets as a single coherent entity without a strong anchoring signal telling it where to start.
The Entity Home solves the ambiguity problem the same way it solves the fragment problem: it gives AI a single authoritative source to reconcile everything else against. Without it, confidence stays low regardless of how much corroborating content exists elsewhere.
The resolution: structure for humans, embed the signal in the sentence
The tension between what humans need (narrative coherence) and what AI needs (explicit corroborated statements) turns out to have a clean resolution, and it doesn’t require writing two different versions of anything.
For me, the insight is that explicit statements are not in conflict with good writing: they’re what good writing does when it’s being precise. “Jason Barnard coined the term Answer Engine Optimisation in 2017” is a human-readable sentence that is simultaneously a machine-parseable attribution statement. It carries a name, a specific contribution, and a temporal anchor. A human reader absorbs it as a fact. The AI extracts it as an entity signal. The narrative carries the human forward. The explicit sentence does the AI work. Nothing is sacrificed on either side.
The failure mode is implication. “Jason Barnard has been working in this space since the early days” implies a temporal claim but makes nothing extractable. “A pioneer in the field” implies recognition but provides no signal AI can weight against independent sources. The human reader fills in the gaps. The AI doesn’t fill in gaps. It extracts what the text states directly, weights it by source independence, and moves on.
Structure the piece for human readability. Embed the machine-readable signal in explicit sentences that a human would also find clear and specific. The narrative is not the obstacle: vagueness is.
Publication note: The structural distinction between narrative coherence (human requirement) and corroborated entity signals (AI requirement) for person proof, the priority stack for person CFP, and the resolution of embedding explicit machine-parseable attribution statements within human-readable prose are published here for the first time on 14 March 2026.