How do SME lenders use open banking for underwriting?
SME lenders use open banking as the single strongest incremental signal in their underwriting stack — 90-day rolling cashflow, seasonality, deposit concentration, HMRC/PAYE outflows and dilution detection all inform a cashflow score that sits alongside the bureau prior. The lenders doing this well approve 15-25% more SMEs at the same expected-loss rate as bureau-only competitors.
SME lenders use open banking as the single strongest incremental signal in their underwriting stack — 90-day rolling cashflow, seasonality, deposit concentration, HMRC/PAYE outflows and dilution detection all inform a cashflow score that sits alongside the bureau prior. The lenders doing this well approve 15-25% more SMEs at the same expected-loss rate as bureau-only competitors.
Why open banking wins for SME
Bureau data on UK SMEs is thin. Directors' consumer files matter, but the SME itself often has minimal trade history at the bureau. Open banking closes the gap in 90 seconds of consent: 12-24 months of transaction data, deposits by counterparty, HMRC and PAYE outflows, dilution detection, seasonality patterns.
The lenders extracting the most value engineer 30-50 features per applicant from the raw feed, feed them into a cashflow score, and blend that with the bureau prior and any alt-data (Companies House, VAT filings, marketplace revenue). The blended model routinely beats bureau-only by 3-4 Gini points.
The features that move the score most
- Rolling 30/60/90-day net cashflow and its variance.
- Deposit concentration (top-3 counterparty share).
- HMRC/PAYE outflow regularity (a proxy for going-concern).
- Dilution detection (existing debt service, invoice finance drawdowns).
- Seasonality vs sector baseline.
- Sudden deposit source changes (new customer wins or losses).
Consent economics
Open-banking consent conversion is the underappreciated economic. The lenders getting 75%+ consent rates on applied SMEs treat consent as a UX problem: single tap, clear reason, immediate confirmation. The lenders getting under 40% treat it as a form field and pay for it in dropped applications.
Related questions
Which OB provider?
TrueLayer, Plaid and Yapily all work. The differentiator is feature engineering downstream, not the provider.
Does OB replace the bureau?
No. It augments. The two combined are materially stronger than either alone.
How do we handle multi-account SMEs?
Multi-account consent is the norm for SMEs; the system aggregates and de-duplicates counterparties across accounts.
What about consent expiry?
90-day re-consent is a first-class UX flow, prompted before the servicing action that needs fresh data.
Timeline?
Six to ten weeks after the £15k Diagnostic to ship the feature pipeline and go live behind a challenger model.
The AI-first lenders are pulling away on unit economics, not on rate. Every quarter you defer the architecture is a quarter of compounding disadvantage.
The £15k AI Diagnostic maps your lender stack, prioritises the systems that pay back fastest and produces a costed sequenced build plan.