Trust Centre
AI disclosures
The system is AI, and says so
EliAI is an Agentic AI Operating System. Its Digital Employees are AI specialists; they are not licensed professionals and we do not represent them as such. Where a communication we send contains content generated or substantially generated by AI, we disclose that.
The underwriter decides
Every output is decision support. Funding, pricing and purchase decisions are made by the people accountable for them, and the system is built to make that possible rather than to make it look unnecessary: findings are attributed to the specialist that reached them, disagreement between specialists is named rather than averaged away, and every figure carries the retrieval that produced it.
It refuses to price what the evidence will not carry
Where the comparable cohort is too thin, a mark is recorded as indicative, carries no band, recommends no advance, and is excluded from book value rather than assumed to zero. The database refuses to store a priced mark without a band, so the refusal is structural rather than a matter of anyone remembering. See the worked valuation for the same file priced and declined.
No customer data trains anything
There is no training and no fine-tuning in this product. The reasoning is grounded in retrieval over a corpus of comparable settled matters, and the corpus and engine version are pinned to each valuation artifact so a mark can be re-derived rather than merely retrieved.
The engine is not named
The compound architecture behind EliAI is Eve-Finance F5/reasoner — five cooperating models composed per request. Which foundation models sit inside it is not published, on this page or any other. That is a deliberate boundary, not an omission: the composition is the engineering, and naming its parts would describe the product as a wrapper around something else.
SecurityData handlingSubprocessorsAvailabilityIncident responseCompliance