Citations that go nowhere
Reference URLs that were never real, or no longer resolve to the page they claim to cite.
CiteAbility™ builds AI information-integrity infrastructure
Building the integrity layer that helps artificial intelligence decide what to rely on and cite.
AI can produce fluent, confident answers while leaning on material that is weak, irrelevant, fabricated, outdated or poorly attributed. CiteAbility™ adds an independent integrity layer around the information and citations used to construct those answers. It does not replace foundation models — it assesses the evidence they stand on.
Citation Integrity™ helps determine whether the information behind an AI answer deserves to be relied on and cited.
Source
Claim
Evidence
Decision
The problem
A plausible answer is not the same as sound evidence.
AI systems can return answers whose supporting evidence has integrity failures even when the wording sounds authoritative and citations are present. These failures are specific, recurring and measurable.
Sources that don't support the claim
Genuine, working citations attached to statements the cited page never actually makes.
Circular sourcing
Dozens of pages appearing to corroborate each other while repeating one unsupported original claim.
The framework
The Citation Integrity™ framework.
A seven-dimension framework for evaluating whether information is dependable, evidenced and attributable enough to support an AI-generated answer.
01
Source Authority
Is the source itself credible and appropriate for the subject?
02
Entity Authority
Is the person or organisation behind the information qualified and dependable for this claim?
03
Evidence & Citations
Can the claims be substantiated, and do cited sources actually support them?
04
First-Hand Experience
Does the source show direct knowledge, testing or original evidence rather than mere repetition?
05
Content Quality
Is the information accurate, complete, specific, current and appropriately qualified?
06
Technical Accessibility
Can an AI system reliably discover, parse and attribute the information?
Integrity Analysis
The final reasoning layer weighs every dimension together and resolves to a clear decision: Cite, Cite with Qualification, Verify, or Do Not Cite.
Who it's for
Two sides of the same integrity problem.
Some organisations need to know whether the evidence behind an AI answer holds up. Others need their own expertise to be discovered, verified and cited correctly.
Builders of AI systems
Know whether the sources behind your answer actually support it.
For AI tool companies, RAG and enterprise AI platforms, and foundation-model and search providers that need an independent read on evidence quality before an answer reaches a user.
Owners of information
Make your information easier for AI systems to verify, rely on, and cite.
For publishers, brands and enterprises whose expertise deserves to be surfaced — and who want the evidence, attribution and technical access to hold up under scrutiny.
Responsible claims
We are precise about what this does.
Citation Integrity™ is designed to reduce measurable integrity failures in AI-sourced evidence. We don't claim to eliminate AI hallucinations or guarantee that cited content is true — we validate specific, measurable improvements in citation and evidence quality.
About CiteAbility™
Founded at the intersection of verification and applied AI.
CiteAbility™ was founded by Chris Emmins and Mark Barclay, combining two decades of information-integrity practice with hands-on AI systems delivery.
Contact
Talk to CiteAbility™ about your AI integrity needs.
Tell us what you're building or publishing, and we'll come back with the most useful starting point.
We work with AI platforms that need to defend the evidence behind their answers, and with publishers and enterprises that want their information to stand up to automated verification.
Independent of any foundation model provider
One framework across sources, pages, claims and answers
Built for regulated and reputation-sensitive environments