Notes on responsible AI in banking.

24 May 2026 · 8 min

What an AI audit looks like in 2026

What a 2026 AI audit examines, how it samples, how findings are categorised, and what to do before one arrives. Written from the perspective of the team being audited.

23 May 2026 · 7 min

How to write an AI acceptable-use policy

The seven sections every AI acceptable-use policy needs, why it stands alone from the general AUP, and the review cadence that keeps it from rotting.

22 May 2026 · 8 min

The AI steering committee: charter, membership, cadence

The standing body that decides what AI is allowed to do inside the company. The charter we drafted before building anything: membership, cadence, agenda, and the four artefacts it owns.

21 May 2026 · 7 min

AI in payments compliance, a regulator-readable framework

The four documents that make an AI compliance framework legible to a regulator without a tour guide, plus the thirty-minute walkthrough we use as a stress test.

20 May 2026 · 7 min

The AI use-case register pattern

What goes in each entry, who keeps it, and why we built the register before we built the Brain.

12 May 2026 · 7 min

Vendor due diligence for AI models

Eight items most AI vendor reviews miss, the contract terms that make them enforceable, and why we publish the list.

05 May 2026 · 8 min

The audit trail for AI banking decisions

The append-only log that records every AI-assisted action in the bank: what is in each entry, how it is hashed, and how a regulator would query it.

28 Apr 2026 · 7 min

Alert triage with AI: cutting noise without cutting signal.

How banks can use AI to triage transaction-monitoring alerts without letting the model close them. First-read precision, tail recall, and the line we hold.

25 Apr 2026 · 7 min

AI-assisted KYC review, what the model can do

The line we draw between what the model does inside a KYC file review and what stays with the analyst, plus the three numbers we measure.

22 Apr 2026 · 6 min

Responsible AI is a list of gates, not a brand statement.

Six gates that every AI workflow inside a bank must clear before it touches money, why each one exists, and how we keep them cheap enough to clear in a day.

18 Mar 2026 · 5 min

Why every Brain answer ships with a citation.

How source attribution works inside Veetso Brain: the four layers that have to cooperate, why we made citations a default, and what drafting mode does instead.

04 Feb 2026 · 4 min

Governance arrives before the technology does.

Why we sequenced our AI roadmap as governance, then knowledge, then intelligence, and what running it in reverse would have cost.