Industry

Financial Services

Production AI for fintechs, banks, and payment companies.

We built the first production version of Volopay, a YC S20 fintech that's now raised $30M. We know production-grade financial software and where AI actually fits in regulated workflows.

Where AI fits in financial services.

Six application areas where we ship production AI in this industry, chosen against the operational realities, not the buzzwords.

01

Transaction reconciliation

AI agents that match payments to invoices, flag anomalies, route exceptions. Live in production for Volopay.

02

KYC & document review

Identity verification, document review, PEP and sanctions screening with full audit trails.

03

Fraud detection

Real-time monitoring with AI-assisted anomaly scoring. Reduce false positives without missing real fraud.

04

Customer onboarding automation

End-to-end onboarding workflows: application intake, identity checks, document collection, underwriting hand-off, with regulator-friendly audit trails.

05

Treasury & spend operations

Spend categorisation, accrual generation, multi-entity reconciliation, FX exception handling. Built for finance teams who've outgrown spreadsheets but can't justify a full ERP migration.

06

Regulated chat & advisory copilots

Suitability-aware customer-facing assistants with disclosure logic, role-based access, and per-jurisdiction guardrails. Human in the loop on every regulated recommendation.

FAQ

Have you shipped AI in fintech production?

Yes. We were the founding engineering team for Volopay (YC S20, $30M raised). AI reconciliation and KYC agents in production across 6 countries. We're also building FLIN's fintech app for Indonesia + Philippines markets.

How do you handle regulatory review?

We document data flows, access patterns, and AI decision logs in formats your compliance team can audit. For regulated AI (KYC, fraud), we build mandatory human-in-the-loop checkpoints.

Can you deploy AI on-premise?

Yes. On-premise, confidential computing environments, or customer-owned cloud. For sensitive data we can use locally-hosted AI models.

What does a typical fintech engagement look like?

Most fintech engagements run 12-16 weeks for a single AI workflow: KYC, fraud, reconciliation, or onboarding, with compliance reviews built into each milestone. Larger multi-workflow programmes run 4-6 months across two or three production releases.

Tell us what you're solving for.

We'll listen first, ask the right questions, and follow up with a clear proposal.