The full story
Details for the curious
Open any section for the full detail — figures, design decisions and how it fits your business.
Document AIRules-based scoringRAG with citationsVerification layer
The challenge
By hand, one credit file takes about 8 hours: roughly 5 spreading statements and 3 writing the memo. Banks want AI that helps the analyst — never one that replaces them.
What we built
- Credit pre-screening assistant: reads the borrower's documents, writes the analysis and recommends a next step, with every finding linked to evidence.
- Rating engine: finds the statements, reads the tables and scores credit metrics from 1 to 7. AI extracts; scoring is rules-based, so the same filing always gives the same rating.
- Cited Q&A: ask a question about a borrower's report and get the answer with the page it came from.
CollectApplication, financials, statements
ReadDocuments become structured data
AnalyseCash flow and policy checks
VerifyEvery figure re-run and traced
Analyst decidesEdits, overrides, signs off
Built so a credit committee can trust it
- The analyst decides — no automatic approvals or declines.
- Every figure traced to its source, calculation and check.
- Runs in your environment: your cloud, your servers or fully offline.
Hours saved use external benchmarks (McKinsey & IACPM: 20–60% analyst productivity gains). We measure your own figure in a proof of concept.
