At a recent EngX leadership dinner in London, senior engineering leaders gathered for a candid conversation about one of the most transformative shifts in modern software delivery: the growing influence of AI and the responsibility that comes with it.

AI’s Expanding Role — and Its Risks

The group kicked off with a discussion on how AI is reshaping engineering workflows. While many attendees expressed excitement about the potential to simplify processes and accelerate delivery, there was an equally strong focus on the risks that accompany AI-generated code.

A consistent theme emerged: trust. Leaders emphasized the importance of building trust in AI systems through strong controls, scalable governance, and mechanisms that help teams understand why an AI system is making certain decisions. Security risks were top of mind as well, especially those created by large dependency injections and new, novel vulnerabilities introduced by AI-generated coding patterns.

Context Is King for AI Agents

One area of strong agreement: AI agents can only be effective when operating with the right context — and that context must be dynamic. Attendees noted the need for agents to stay current with changing environments, updated designs, merged PRs, and active deployments. Static context simply isn’t enough in fast-moving engineering organizations.

Operationalizing Governance and Compliance

The second half of the evening focused on how organizations can operationalize governance and compliance without slowing down delivery. Leaders discussed:

• Integrating governance into the SDLC, especially through tighter alignment with change control systems.

• Shifting CABs (Change Advisory Boards) left — approving what teams intend to build and deploy, rather than reviewing work already completed.

• Embedding risk mitigation into everyday workflows, automating wherever possible to reduce manual checkpoints while maintaining strong oversight.

• Reinforcing accountability, where engineering teams own the quality and safety of their deployments just as much as change management functions do.

Culture as the Foundation

Across every conversation, one message stood out: technology alone isn’t enough. Organizational culture — how teams communicate, collaborate, and take ownership — plays a defining role in delivering software safely and effectively. Governance succeeds best when paired with a culture that values transparency, responsibility, and continuous improvement.

We’re grateful to every EngX member who joined this dinner and contributed to such an open, thoughtful discussion. These conversations are what make the community special, and we’re excited to continue exploring how AI, governance, and culture come together to shape the future of software delivery.