Infosecurity Europe: Practical Lessons From Lloyds' Agentic AI Security Playbook

Lloyds Banking Group presented a pragmatic, engineering‑focused ‘AI Safe Adoption’ strategy that centralizes agent registration in an internal marketplace, enforces multidisciplinary governance and production gating, pilots multi‑vendor agent identity models, constrains agent capabilities to reduce blast radius, and scales assurance through runtime observability and automated red‑teaming (including OWASP Top 10 for Agentic). The report emphasizes starting with low‑risk, high‑value use cases, codifying automated controls, and hands‑on adversarial testing while noting measured business value and planned expansion of GenAI use cases.