Inside HL Hunt’s Hunt Score: credit decisions in under 250 milliseconds
The AI Underwriting platform scores more than 1,000 data points per application and says it approves more borrowers at a lower default rate than bureau-only models.
The gap between a completed loan application and a funded loan is, for most lenders, a manual underwriting queue measured in hours or days. HL Hunt Financial's AI Underwriting platform, which it markets around a proprietary risk score it calls Hunt Score, is built to collapse that gap to a fraction of a second.
According to the company's published materials, the platform analyzes more than 1,000 traditional and alternative data signals per application — bureau data such as FICO and VantageScore alongside bank transaction history, rent payments and utility records — and returns an approve, decline or refer-to-manual-review decision in under 250 milliseconds, delivered through a REST API and configurable as a decision waterfall a lender sets itself.
HL Hunt cites specific performance figures for the platform: roughly 34% more approvals and default rates more than 50% lower than what it describes as traditional bureau-only scoring, alongside a stated fraud-detection rate of 99.7% of fraudulent applications caught before approval. These are the company's own published figures and have not been independently verified by FinCrunch.
The alternative-data layer is pitched at the same thin-file problem HL Hunt's Business Credit Builder addresses from the borrower's side, approached here from the lender's side instead: an applicant with no bureau file is not, under this model, automatically an applicant with no score, because rent, utility and bank-transaction history can substitute for a missing credit history.
On model governance, the company states the platform generates adverse action notices automatically, is built to ECOA and FCRA compliance requirements, and runs continuous disparate-impact monitoring. It also describes the underlying models as self-learning, retraining over time on a lender's own outcomes data rather than shipping as a static model.
Pricing runs in three tiers: Starter at $495 per month plus $2.50 per underwriting assignment, Professional at $1,495 per month plus $1.75, and Enterprise at $4,995 per month plus $0.95, with the per-assignment rate falling as a lender's volume rises. HL Hunt lists coverage across more than 55 credit, insurance and eligibility products — from buy-now-pay-later to mortgage pre-qualification to business equipment financing — offered white-label, so a lender's own borrowers never see the HL Hunt brand.
A model that continuously retrains on a lender's own outcomes is also a model whose behavior can drift without an independent audit catching it early — the standard due-diligence question for any institution evaluating the platform is not whether disparate-impact monitoring exists as a stated feature, but how it performs in production, and what validation data HL Hunt can provide specific to a prospective client's own borrower population and product mix.