Generative AI is already reshaping how organisations build software, deliver change and operate day-to-day. Adoption is happening across teams, often cutting across traditional governance, security and delivery models.
While the risks themselves aren’t entirely new, GenAI has changed how they surface. Systems interpret intent rather than execute fixed rules, outputs are non-deterministic, and meaningful change can happen outside normal release cycles.
This playbook, written by Chris Rutter (Global Security Lead), Ben Wilkes (AI Engineering Lead) and Phil Parker (Global Head of Technology Strategy & AI Delivery), sets out practical principles and decision frameworks to help you adopt generative AI safely and confidently. It focuses on durable principles rather than fear or fast-moving threat lists, and enables progress without compromising control.
In this playbook you’ll find:
When you download the playbook, we’ll also send you our AI security foundations benchmark — a companion resource to help you assess your current approach, understand what good looks like in practice, and decide where to focus next.
We’re continually developing new resources for leaders and practitioners working with AI, and by signing up you’ll receive occasional updates when relevant new thinking and tools are available.
If you’re responsible for balancing innovation, delivery speed, and security, this playbook is for you.
We help leaders deploy secure, practical AI approaches within real delivery environments.
Alongside the playbook and AI Security Foundations Benchmark, we run focused working sessions to explore relevant AI attack patterns, review practical controls, assess your current approach against the benchmark, and agree clear, proportionate next steps.