For years, the job of an engineering manager or director was clear. Own the project. Run the program. Build the vision. Align people around it. Deliver.

The manager's value was coordination: turning strategy into plans, plans into teams, teams into results. Technical depth helped, but it was not the core of the role. Many directors had not written production code in years. That was normal. Nobody expected it.

That model is ending.

What AI brought

AI did not arrive as one tool. It arrived inside every tool we already used.

Jira drafts tickets and breaks epics into tasks. Spreadsheets build models from a sentence. Chat tools summarize threads you did not read. Meeting software transcribes the call and proposes next steps before you close the window. Code review bots comment on pull requests before a human opens them. Agents write, test and ship features.

The coordination layer, the work managers were paid for, is now partly automated. Status reports write themselves. Follow-ups write themselves. Task breakdowns write themselves.

So the question every company is asking: what is the manager for now?

What companies expect now

I talk with many engineering leaders and companies. One message comes up again and again: managers and directors must be hands-on with AI.

Not "aware of it." Not "supportive of experiments." Hands-on.

They expect a director to:

  • understand how AI tools actually work and where they fail
  • use them daily in their own work
  • ship features with AI, not only approve them
  • lead AI adoption across their org, with real results

A leader who cannot do this cannot judge what their teams are doing. They cannot tell a real productivity gain from a demo. They cannot see where AI adds risk. They cannot set realistic goals, because the baseline of "what a team can deliver" has moved, and they do not know where it moved to.

From alignment to production

This shifts the role. Management moves from alignment and vision toward production-embedded work.

The manager is back in the build. Not as the best coder on the team: that is not the point. The point is shipping outcomes. A director who can take a feature from idea to production with AI tools understands the real cost, speed and risk of that work. That understanding shapes better decisions about scope, staffing and strategy.

Craft still matters. But the measure changes: speed to working outcome, with quality owned and verified, not hand-built.

Vision does not disappear. It changes shape. Company goals now depend on how well the org uses AI. A director who does not understand the tools cannot write a credible vision for a company that runs on them.

What this means

For managers: the safe middle ground is gone. Pure coordination is being automated. Your value is judgment, and judgment needs hands-on experience.

For companies: hire leaders who build. Ask them to show, not describe, how they use AI.

For teams: expect your manager to work next to you, not above you.

The manager of 2020 aligned people. The manager of 2026 aligns people, and ships.