Smart Intake · pilot proposal

Loan decisions in under one hour,
compliance preserved.

Finance House currently quotes a 4-day initial response on loan applications. The EU consumer-lending market has moved on — Revolut, Klarna and the neobanks underwrite in minutes. This is what an MFSA-compatible AI-assisted intake layer could do on top of Finance House's existing process, without touching the regulatory framework.

Hero loop · 10s · restrained / UI-realistic

[ Higgsfield product walkthrough renders here ]

The gap we found

Finance House's own FAQ describes the bottleneck precisely.

From financehouse.mt/help/faqs

"We endeavour to provide an initial reply within 4 days (excluding Sundays and public holidays) from submission of loan application subject to having all the correct details and paperwork."

That's faster than BoV or APS — and your testimonials reflect that win. But Revolut's car-loan partners underwrite in under 4 minutes. The competitive ceiling has moved. The question is whether Finance House moves with it or holds the line.

Current

Manual review

·Customer uploads 6+ documents

·Officer reviews payslips / FS3 by hand

·Manual KYC + AML/CFT screening

·Credit + affordability review

SLA: 4 days

FH Smart Intake

AI-assisted intake

✓OCR auto-extracts all document data

✓Live KYC + AML pre-screen

✓Affordability auto-calculated

✓Officer reviews ONLY exceptions

SLA: <1 hour clean / 24h exceptions

Compliance-preserving by design

Every AI decision logged with full reasoning chain, ready for MFSA audit. Human officer retains final approval authority — AI is a pre-screen and accelerant, not a replacement for the regulated decision-maker.

Try it · partial preview

FH Smart Intake — live demo

A working slice. Upload a sample document or use the demo data — watch the AI extract, verify, and decision in real time.

Smart Intake · Application Review

live

Step 1 · Documents

Step 2 · AI extraction & decision

Demo runs locally · Full version integrates with FH's existing loan origination system + ComplyAdvantage / Onfido / OpenAI vision OCR

Watch · 25 seconds

4 days to 47 minutes — without breaking compliance.

Problem → solution narrative · 25s · UI-realistic

[ Higgsfield narrative renders here ]

The full build

What this becomes once we wire it into Finance House's stack.

01 · Maltese ID / passport OCR

Tuned for Maltese ID-card format + EU passports. Face-match liveness for non-Maltese applicants. Verified residency proof workflow.

02 · FS3 + payslip extraction

Income parsed from FS3 + last 3 payslips. Self-employed: P&L + Tax Statements ingested. Affordability calculated against your existing risk model.

03 · Live KYC + AML/CFT pre-screen

EU sanctions, PEP, adverse media. ComplyAdvantage or Onfido — your existing provider if you have one. Risk score attached to every application.

04 · MFSA audit-ready logging

Every AI decision recorded with full reasoning trace + confidence scores. Regulator can audit any decision back to its inputs. Human officer signs final approval.

05 · Conservative envelope, expanding

Phase 1: AI handles only the easiest 30% of applications (clean payslip, Maltese national, employed). Officer reviews everything else. Envelope expands as confidence proves out.

06 · Maltese + English UX

Customer-facing flow speaks both languages natively. Tested with John Busuttil-style applicants (your existing customer base, your existing language).

If this is on your roadmap

Reply "show me the full one" and I'll send the implementation plan.

Six-week phased rollout: OCR pipeline build, KYC provider integration, MFSA compliance review of the AI decision logging, pilot on a conservative application tranche, expand envelope as confidence proves out.

No follow-up if it's not relevant. Genuinely.