Layer 03. Calls, email and meetings captured without asking anyone to write them up. This is what the bank has heard — and it is what keeps contact recency current, which is what conviction is scored against.
Add a line the client said. The report on the right updates immediately — that loop is the whole point of the layer.
Layer 05. Peer-benchmarked cross-sell, priced in pounds and ranked by conviction. Anything the client raised themselves leads — and is labelled, because their own words and a peer benchmark are different kinds of evidence.
Everything above is already matched to your business. No application form, no document pack to assemble.
We use what the bank already holds. You confirm it is current — you do not re-enter it.
Underwriting runs against your own figures. Your relationship manager sees the same screen you do.
Not "who is our biggest client" but "where is the largest gap we can justify". Headroom is the total opportunity; risk-adjusted discounts it by conviction. Where the two are far apart the opportunity is large and the evidence is thin.
Ask a question in plain English. It is translated into a read-only graph query grounded in the lending ontology — never a guess, never a write. Open the live graph viewer →
Layer 02. Screening runs first and can stop everything; classification only runs for an application that may proceed. A cleared client is classified into the taxonomy and enters the peer benchmark the same day.
Legal name
Business description
Annual turnover (£)
Directors & owners (comma separated)