Back to Posts
What the evolving Engineering Manager role looks like in practice?

What the evolving Engineering Manager role looks like in practice?

~150 engineers. AI-powered DevX tooling rolled out. Executive support. Three months later: only 30 of 150 using the tools. At AWS Summit London yesterday, one talk got me.

Not the engineers ignoring the tool. The manager who thought making them available was enough.

The Engineering Manager role is changing in ways most job descriptions havenโ€™t caught up with yet.

The role has always lived between player and coach. Technical enough to understand the work. Removed enough to see the bigger picture.

AI is pushing that balance to a new place entirely.

The execution layer has changed. High performers now know how to collaborate with AI, when to trust it, and when to challenge it. That's a skill, and most teams are figuring it out without much guidance.
The coaching dynamic has shifted, too. Junior engineers often adopt AI tools faster than seniors, at first. Which sounds like good news, until you realise the real advantage has moved away from raw output, towards judgment, context, and knowing when the AI is wrong. AI doesnโ€™t teach that. Leadership does.

๐—ฆ๐—ผ ๐˜„๐—ต๐—ฎ๐˜ ๐—ฑ๐—ผ๐—ฒ๐˜€ ๐˜๐—ต๐—ฒ ๐—˜๐—  ๐—ฟ๐—ผ๐—น๐—ฒ ๐—ฎ๐—ฐ๐˜๐˜‚๐—ฎ๐—น๐—น๐˜† ๐—น๐—ผ๐—ผ๐—ธ ๐—น๐—ถ๐—ธ๐—ฒ ๐—ป๐—ผ๐˜„, ๐—ฝ๐—ฟ๐—ฎ๐—ฐ๐˜๐—ถ๐—ฐ๐—ฎ๐—น๐—น๐˜†?

โ€ข Getting in the game yourself. Using the tools. Breaking them. Understanding the failure modes. You lose credibility fast if your team knows the tools better than you do, and they will, quickly.

โ€ข Owning adoption from announcement through execution. Workshops. Real use cases. Weekly check-ins. Champions who bring others along. The talk confirmed what I've seen: ongoing enablement drives adoption, not rollout emails.

โ€ข Designing how the team works with AI as part of the workflow. What context does your AI assistant have about how your team operates? What standards, what patterns, what tribal knowledge? Someone needs to think about that deliberately. Right now, few teams are doing that deliberately.

โ€ข Building judgment capacity through coaching. When AI handles the routine work, the conversations shift. Less "did you finish it" and more "how did you approach it, what did you verify, what did you decide to change." That's a different kind of 1:1.

The EM career path has always rewarded people who can hold technical credibility and people leadership at the same time. That's still true.

But the playing coach model is being stretched. The coaches who are also genuinely on the field, who understand the new game from the inside, are going to look very different from the ones managing from the dugout.