Banking and fintech where every number has someone to explain it

ID, proof of address, source of funds, company deeds. The onboarding flow reads each document, checks that it is legible, current and consistent with the others, and asks for anything missing on the spot.

Sanctions and politically exposed person screening runs through the provider you already use. AI prepares the anti-money-laundering file; the analyst signs off on accepting the customer.

Transactions are read through the bank’s API or open banking (PSD2) and matched against invoices, remittances and payments by amount, date and reference. Whatever matches is reconciled and posted; whatever does not goes to a queue with the difference calculated and the most likely candidates.

A transfer reference is rarely clean. That is where the AI helps: it understands that “INV 2231 PARTIAL” and invoice 2231 are the same thing, and proposes it. A person confirms uncertain cases, and those confirmations become written rules, not a black box.

Payslips, tax returns, annual accounts and statements arrive as PDFs. We read them, extract the figures your credit policy requires and leave the file ready, with every figure linked to the page it came from.

The EU AI Act treats creditworthiness assessment of individuals as high-risk. So AI does not score or decide: it prepares, and the analyst decides with everything in view.

Mobile apps, back offices and APIs for fintech: a double-entry ledger, idempotent payments that never charge twice, reconciliation with the payment gateway and an audit log that cannot be edited.

We integrate gateways such as Stripe, Adyen or Redsys, and document access, suppliers and continuity for what we build, which is what the DORA regulation requires of financial entities.