Example · Engineering Manager
Wider teams are coming whether or not agents work. Be the manager who can run eight people and still read the diff.
35 of your 40 hours still need you
- 21 stay yours
- 14 faster with AI
- 5 tools could take
A typical week at the middle pace. Your plan uses your own tasks.
Are engineering manager jobs disappearing?
By our task-by-task count, 35 of a typical 40-hour week as an engineering manager still need a person. Tools could already take about 5 hours. Big tech has been removing manager layers since 2023, and the reasons given were cost and speed rather than models. Google said in August 2025 it had 35% fewer managers of small teams than a year earlier.
Read more
Amazon set a goal in September 2024 to raise its ratio of individual contributors to managers by at least 15%, then cut 14,000 corporate roles in October 2025 and 16,000 more in January 2026 while "reducing layers."
Most of what makes your seat worth having is the least exposed work you do: people, hiring, delivery calls. About 53% of your week sits there. The risk to you is the count of seats. A manager who can show a well-run wide team, with agents in the workflow and evidence in hand, is the one who keeps a seat when two teams merge.
The hard truth
Manager engagement is falling faster than anyone else's.
Managing people is getting harder on the people doing it. Worldwide, manager engagement dropped five points in a year, while the people they lead held steady.
Sources (2)
- Global manager engagement fell to 22% in 2025 from 27% in 2024. Gallup, State of the Global Workplace 2026, 2026-04
- Individual contributor engagement was 19% in 2025, up from 18%. Gallup, State of the Global Workplace 2026, 2026-04
What stays yours
1:1s, coaching and career growth
This rests on relationships and context. Trust built over months is the input. A model can draft a growth plan, but it cannot know your strongest engineer is interviewing elsewhere because of a family move.
Performance management: calibration, improvement plans, exits
This needs context and the accountability you carry for the employer. Someone must stand behind a rating or a termination in front of HR and sometimes a lawyer.
The hire or no-hire call
This depends on verification. A bad hire is discovered around month six and costs a quarter. The debrief is where your judgment matters most.
Delivery commitments and saying no
The accountability here is informal. Your name sits next to the date. When scope has to be cut, product and leadership want one person who will answer.
Incidents and production risk calls
This needs verification and accountability. Deciding whether to roll back at 2 a.m. takes a person with context and authority.
What's the next role for an engineering manager?
Plan A is your best move inside the work you do now. Most sessions in the 13-week plan build toward it.
- A
Become a senior engineering manager who owns AI delivery for a group
Why it fits
Why you. You already run a team where agents are being adopted, and you sit where output, review and quality meet.
The gap. You don't yet have measured numbers, written rules for agents or a tech lead strong enough to own review.
The trade-off. You are building the case that one manager can run more people. It may be used to widen your span or merge a peer's team.
Other moves that fit
These are also good bets in this field. Read them if Plan A doesn't suit you.
Lead developer AI enablement or productivity
Why it fits
Why you. Postings now ask for a manager who drives agent adoption across teams. Rocket Companies (August 26, 2026) and Money Forward in Tokyo (October 15, 2026, ¥9M to ¥18M) both posted this shape.
The gap. You have run one team. These roles want evidence of adoption across several.
The trade-off. Postings are still few, and the role may fold into platform teams once adoption settles.
Become a manager of managers
Why it fits
Why you. If you already run a wide team well, you are a natural next pick.
The gap. You need a tech lead or new manager you have grown, and a record of running through other people.
The trade-off. This is the layer most often removed. Among LeadDev respondents who saw fewer manager roles, 65% said middle management was hit (LeadDev Engineering Leadership Report, April 2026).
What's a plan B for software engineering managers?
Plan B is a job AI is creating that these skills already fit. One session a week in the plan trains you for it, so it is ready if your job shrinks or you want a change.
BPlan B, picked for this role
Developer AI enablement lead
You already set your team's review rules and own delivery quality, which is this job across several teams.
Other options
AI enablement and adoption lead
You coach engineers through new habits in 1:1s and team rituals, which is how AI adoption actually sticks.
Agent workforce manager
You already set goals, review output and decide who approves what for a team, and agents need the same.
In your own plan you can switch to any of the 47 roles.
The first two weeks
The first 2 of 13 weeks. Each week has three sessions of 30, 45 and 30 minutes, plus one 40-minute Plan B session. By week 13 you'll have an evidence file for each person, a one-page memo asking for a senior manager title, and a developer AI enablement playbook.
- Week 1: Measure your week
What you do
- ATrack one week of your work against the task list on this page. Write down roughly how many hours went to each task.
- ACompare the last two quarters of commitments with what actually shipped.
- AList every decision you made last month. Next to each, write who else could have made it.
- BTake the first half of Anthropic's free AI Fluency course. Write down its four habits for working well with AI.
- Week 2: Start an evidence file
What you do
- AWrite one page per direct report with their wins and misses since the last review cycle. Pull from pull requests, docs and your notes.
- AAsk your director in writing what evidence calibration (the cross-team rating review) weighs most.
- APut a 10-minute weekly reminder on your calendar to add one line per person.
- BFinish the AI Fluency course and pass its final assessment.
Three courses from the plan
Checked 2026-10-15.
- AI Fluency: Framework & FoundationsAnthropic Academy · Free
Yes. It is a fast, free baseline for weeks 1 and 2.
- AI Capabilities and LimitationsAnthropic Academy · Free
Yes. It names the failure types you will see in agent-written code.
- Claude Code in ActionAnthropic Academy · Free
Yes, if your team uses coding agents. It covers permissions and hooks, which become your team rules.
Risks beyond AI: highDelayering risk
Manager seats get cut when companies chase cost and speed, and that started before agents could write much code. When two teams merge, one manager seat goes, and the work that held does not save the person whose team was folded in. Agents make the case for wider spans easier to argue, which is why this risk outranks the AI one for your seat.
Where this could be wrong
- At a startup you may already be a player-coach with four reports, and coding is a much bigger part of your week than we scored.
- The review-load problem may be a 2026 artifact. If AI review tools get good, code review becomes one of the more exposed tasks and the player-coach case weakens.
- Span widening could stall if more companies hit Meta's problem. If that happens, the risk of losing manager layers drops from high to moderate.
- We did not score agents running 1:1 prep or calibration evidence. If employers allow it, performance work becomes more exposed, with AI doing part of it.
Also matched: engineering manager, software engineering manager, software development manager. Research checked 2026-10-15. Page updated .