AI Tools in Regulated Banking: Monzo's Control Model
summary
Monzo runs an opinionated monorepo with 3,000+ microservices and uses static analysis, automated code mods and service generators to enforce auth and dependency patterns; models like Claude and Codex act as engines behind a bespoke LLM gateway that routes calls, falls back between providers (Opus mentioned) and logs telemetry. Background agents execute long jobs via Slack and a central UI but require runtime guards to prevent runaway token costs; measured wins show faster PR velocity, not reduced headcount.
AI-written summary · about
chapters (6)
- 0:00Monzo's platform advantage
- 4:29AI as a new engineering muscle
- 16:10Static analysis and code mods get cheaper
- 20:32Background agents without losing monorepo control
- 28:14Controlled access beats blanket restriction
- 31:39Owning the interface to AI tooling
related talks
24mAI Governance in Financial Services: What Every Developer Should Know
33mOptimizing Bank Operations: Design Patterns, NFR, and GenAI Fusion
42mAgentes en el monorepo
01h 04mPanel Discussion: Money Moves at Machine Speed - Trust, Risk, and Quality in Agentic Finance
21mBuilding a Neobank With Agentic Engineering
15mThe 85% Beyond the Model
45mAI agents in practice - beyond the prototype
49mPanel Discussion: Driving the Agentic Shift in Banking Adoption Governance and Scale