It is one of the oldest questions in business. Which job is the hardest?
One of the most common answers is the manager in the middle: accountable to the number, responsible for the people, and holding the space between the decisions senior leadership makes and the individual contributors those decisions land on.
If that is the most common answer, we owe those managers a great response to the AI era. They are the link between the people doing the work and the mission of the company.
So far, most of the investment has gone somewhere else.
The individual contributor got tools. Research assistants, call prep, drafting, summarization, outreach. Real ones, and they work.
Senior leadership got visibility. New reporting, forecast modeling, scenario analysis. Dashboards that answer questions nobody could answer two years ago.
The middle got expectations. They were handed the rollout, told to drive adoption, asked to produce more with the same headcount, and given a number that assumes the tools are already working. What they were not given is anything built for the job they actually do.
That is not a tooling gap. It is a design gap.
Why it breaks.
A manager is a transmission. Strategy arrives from above, capability sits below, and the manager converts one into the other. If the transmission does not change, nothing downstream changes either, no matter what you install at both ends.
Consider what a rollout actually asks of them. Learn the tool on their own time. Teach it to a team that has questions they cannot answer yet. Hold the number steady while everyone is slower for a month. Then report adoption upward as though people logging in were the point. None of that work appears in the plan, and all of it lands on one person.
Give a manager AI-generated insight and no time to act on it and you have not helped. You have added an input. It is the same pattern we have written about whenever tools get layered onto an operating model nobody redesigned.
There is a quieter cost too. Middle managers are asked to sponsor a change they were not trained for, in front of a team that watches how they handle it. When a rollout lands badly, they absorb the credibility hit. That is a real thing to ask of someone, and it is worth saying out loud.
What the layer above owes them.
Setting an expectation costs nothing. Enablement is the actual work, and it has three parts.
Teach the job, not the tool. Training a team on software is not the same as showing a manager how the week changes. The second is harder, and it is the only one that transfers.
Make room to invent. The best manager workflows will be built by managers, not handed down in a rollout deck. That takes time which is not already spoken for, and permission to try something in front of a team that might not work the first time. Creativity is not a personality trait here. It is a function of whether anyone has the room for it.
Then get out of the way. The instinct after a rollout is to standardize immediately. Standardize too early and you lock in the first idea anybody had. Let a few managers build something better, then make that the standard.
Enough structure to align, enough freedom to invent. It governs a leadership team the same way it governs a company.
How we think leaders should use it.
To prepare, never to decide. The highest return for a manager is walking into the coaching conversation, the pipeline review, or the hard one-on-one already knowing what to look at. The judgment stays theirs. The preparation should stop costing them the evening.
To see the team, not the dashboard. Reporting tells you what happened. The more useful question is what is happening across people: who is drifting, whose deals stall at the same stage, which behavior is spreading. That is a pattern problem, and patterns are what machines are genuinely good at.
To defend the calendar. Most of a manager’s week is work that does not require a manager. The point of removing it is not speed. It is buying back the hours only they can spend, on the few acts that actually move a team.
No manager can do that alone, because most of what fills the week arrives from above. So the harder question belongs to senior leadership: are we asking too much, have we ranked what we ask for, or are we asking for everything and calling all of it a priority? A manager cannot protect time that the people above them keep spending.
To be argued with. The most underused prompt in management is asking the machine to attack the plan before the team has to live with it.
The standard. If AI is working for your managers, it shows up as time returned and decisions sharpened. If it shows up as one more thing to log into, it was built for someone else.
Nobody builds for the middle by accident. It has to be designed for them, on purpose, around the week they actually have.