AI has made a lot of people faster.
It has not made them better.
That is the uncomfortable bit. A strong engineer with AI tools can prototype faster, test more ideas, write more code, clean up more edge cases, and compound their judgement across more work.
A weak engineer with AI tools can also move faster.
They just create more mess.
The same is true for product managers, designers, marketers, operators, and leaders. AI gives everyone leverage. But leverage does not care whether the person holding it has good judgement. It just multiplies what is already there.
AI makes bad direction more expensive
There is a lazy version of the AI conversation that says managers matter less now.
I think the opposite is true.
Managers matter more because the cost of unclear direction has gone up. When work was slower, bad direction had a natural speed limit. You could catch mistakes in planning, design, review, or delivery because the whole process took longer.
Now people can build the wrong thing much faster and that changes the job.
The question is not “did we make more things?” The question is “did the work move the business?”
That distinction matters more now because the volume of plausible-looking work is going to explode. Teams can ship features, write documents, create dashboards, generate customer research, automate workflows, and still miss the point.
A bad plan with ten times more execution is not a better plan.
It is just a more expensive mistake.
This is why goal clarity becomes non-negotiable. A fuzzy direction that used to waste two weeks can now waste two days and still produce enough output to look impressive.
Speed does not fix unclear priorities.
It exposes them.
If the goal is vague, AI just helps the team wander faster.
Managers need to build
A manager cannot sit above the work, vaguely steer, and wait for updates. They need to understand how the work is actually being done. They need to know what AI tools make possible, where they create risk, and what good output looks like now.
If you manage a team using AI but never use the tools yourself, you are managing through folklore.
You are guessing.
There was a version of management where the job slowly became coordination. Meetings, updates, stakeholders, roadmaps, escalations, hiring plans, performance conversations. All of that still matters, but AI has changed the tradeoff.
There are now moments where the fastest way for a manager to help is not to call a meeting, write a doc, or protect the team from a request.
It is to build the first version.
Not because managers should become full-time individual contributors again. Not because the team needs a hero. But because building is often the shortest path to clarity.
A prototype can end a circular conversation. A draft can unblock a decision. A quick analysis can replace three meetings about whether analysis is needed.
Managers who cannot build will increasingly manage from too far away.
The bar has moved
Managers should expect more.
That will make some people uncomfortable, but it is true.
If a team has access to tools that remove hours of grunt work, then the old pace cannot be the default expectation. Small improvements should happen more often. Documentation should be less neglected. Basic research should be easier. First drafts should arrive faster. Repetitive tasks should not sit around waiting for someone to find time.
“I did not have time” means something different when the tools can do a lot of the first pass.
But raising expectations does not mean asking everyone to produce more noise. More code, more documents, more prototypes, more research, and more tickets can still be waste if they are pointed in the wrong direction.
The new bar is not more output, it is better use of leverage.
A few expectations should change:
-
Small useful work should keep moving. Bugs, documentation, cleanup, small product improvements, basic tests, internal tooling. These should not sit untouched for weeks because everyone is “focused on the big thing.”
-
Output needs to be tied to outcomes. More work is not automatically better work. A person who creates ten documents nobody uses has not beaten the person who created one decision. A team that ships more but does not move the metric is not obviously doing better.
-
Test coverage is more important. If the team can write code faster, then the cost of bugs goes up. If the team can ship faster, then the cost of bad decisions goes up. Tests are a way to catch both before they hit customers.
-
Partial work is less acceptable. Not because everything needs to be perfect, but because the first pass is cheaper now. Competitive research, ticket analysis, customer summaries, draft proposals, basic data analysis. If the tool can do the rough version, turning up with nothing is harder to defend.
That means people need stronger judgement, clearer ownership, and better taste. They need to know when AI has helped and when it has produced something that only looks useful. They need to know when to ship, when to edit, when to ask for help, and when to stop.
AI makes weak work look stronger for longer. A bad hire with powerful tools can create more code to review, more decisions to unwind, more documents to ignore, and more false confidence in the room.
That is why the hiring bar also moves.
You need people with enough depth to know when the machine is giving them nonsense.
Collaboration gets harder
AI makes individual work easier, which sounds good until everyone disappears into their own toolchain.
One person has agents writing code. Another has a research workflow. Another has a pile of generated docs. Another has prototypes, summaries, experiments, and half-finished ideas moving in parallel.
The team gets faster, the product gets less coherent.
This is where management matters. Someone has to keep pulling the work back together. Someone has to ask whether the pieces still fit. Someone has to make sure the team is not just producing five different versions of progress.
AI increases throughput. It does not automatically create alignment.
The real lesson
AI will not save bad management, it’ll make bad management louder.
Unclear goals will create more wasted output. Weak performers will create more mess. Managers who do not understand the tools will set bad expectations. Teams without alignment will ship faster and feel less coherent.
But the opposite is also true.
Good managers now have more leverage than ever. They can build faster, clarify faster, test faster, and raise the standard of what good work looks like.
The job is not going away.
The bar is going up.