AI transformation in enterprise — BT Group × Product School

What the video covers

Kerry Small, COO at BT Group, sits down at ProductCon London to discuss how a company that predates the telephone itself is transforming into a customer-centric, AI-native product organization. The conversation covers BT’s three-layer AI stack, the role of guardrails in enabling speed, and how the team evolved its product management practice to match the pace of AI development.

Who it’s for

Product leaders at large, established enterprises — particularly in telecoms, financial services, or government — who face the dual challenge of modernizing legacy systems and building AI capabilities simultaneously. Also relevant for PMs interested in how enterprise-scale AI transformation differs from startup-style experimentation.

Key takeaways

1. The 360 product manager owns business outcomes, not feature delivery. BT shifted its PM role from delivery agents who ship roadmaps to business owners responsible for revenue, gross margin, and customer outcomes. Small calls this the “360 PM” — someone who understands both the technical implementation and the commercial impact of their decisions, which becomes essential when AI capabilities introduce new cost-revenue tradeoffs.

2. Full stack transformation is non-negotiable. Jumping straight to AI applications without addressing infrastructure and data is a recipe for failure. BT’s approach works in three layers: re-architecting core network infrastructure, treating data as a product, and then building AI applications on top. Small argues that organizations that skip the bottom two layers find their AI efforts constantly blocked by foundational issues.

3. Guardrails enable speed, not slow it down. BT built secure-by-design platforms that allow product squads to experiment and fail fast without creating enterprise-level risk. The counterintuitive insight is that teams move faster when the platform handles compliance, security, and monitoring automatically, because they spend less time on approvals and more time on iteration.

4. The 20 percent growth rule reshapes teams. BT strategically allocates star performers and AI-native hires to a dedicated growth portfolio while aggressively retiring legacy technical debt. This creates a clear organizational signal about where the company is headed and prevents transformation energy from being diluted across maintenance work.

5. PM and engineering roles are converging into co-creation. The modern PM must be “the engineer’s best friend” — not by writing code, but by providing such precise outcome clarity that engineers can move autonomously. Small describes this as productive tension: PMs push on outcomes, engineers push on feasibility, and the overlap is where AI products actually ship.

Worth watching if…

You lead product at a large enterprise and need practical patterns for restructuring your teams and technology stack around AI without abandoning the operational stability your business depends on.