Stop Thinking About AI as a Tool. It's Talent.
Over the last few weeks I oversaw teams doing several projects: an advanced Monte Carlo data simulation, an advertising video with six different scenes, a comprehensive health dashboard showing claims and absence outcomes, a detailed presentation about program evaluation, and several well-referenced research papers.
But I have no employees.
Without hiring anyone, I am now a manager.
Suddenly, I have experts.
Here's who showed up when AI arrived: advanced data scientist, research librarian, photographer, videographer, editor, coder, graphic artist, diagnostician, second medical opinion, musician, and more. I also have assistants who take notes, summarize meetings, suggest edits, and send emails.
That's not a tool. It's a team.
And yet most organizations are introducing AI like software — a one-hour tutorial, a login, a password, done. It's not software. It's a team. And teams need something entirely different.
This is why AI rollouts stall. Not just resistance. Not just technical complexity. Not just budget. People are being asked to manage a capable, fast-moving team with no job descriptions, no onboarding, no review process, and no way to estimate what the work will cost. Nobody succeeds at that. Not with people. Not with AI.
And the reason it keeps getting introduced this way? Most of the people making that decision haven't lived it. They've seen the demos. They've read the reports. But they haven't personally managed four AI workstreams at once and felt one go confidently sideways. To them, AI still looks like software — something you click on, like Word or PowerPoint. They don't see what it actually is: turning very smart, capable workers loose with no context, little instruction, and limited oversight.
The thing nobody told you about your new team
Here's what makes this team unlike any other you've worked with.
None of them have skin in the game.
A lawyer who gives bad advice loses their license. An accountant who makes an error is liable. An employee who ships bad work gets a performance review. Your AI team? A mistake will not impact their income, their reputation, or their wellbeing. They are well-intentioned, fast, and genuinely talented. But the consequences of their work flow in one direction only.
Toward you.
This means something that gets almost no attention in the AI adoption conversation: the more capable your team, the more accountable you become. Speed doesn't reduce the need for guidance. It amplifies it. A slow worker who goes off track is correctable. A fast, confident one who goes off track (across four simultaneous workstreams) is a different problem entirely.
Think back to those projects I mentioned doing (the video, analysis, dashboard and papers) at the same time. Those, simultaneously, are vastly more than I could do a year ago. But it means I have to be clear — genuinely clear — about the goals, boundaries, methods, and consistency requirements for all of them, all at once.
Because here's what happens when I'm not.
The work is stunning. Fast, polished, confident. And wrong.
Not broken-wrong. Wrong-and-amazing. A beautifully rendered answer to the question I didn't quite ask. And because it arrived so quickly and looks so good, I'm now further from my goal than when I started.
The manager you didn't know you had to become
When new people join a team, someone has to explain the job. Someone has to review their work. Someone has to be responsible for their output.
That someone is you. Whether you've managed anyone before or not.
Most people who are struggling to get traction with AI aren't failing at technology. They're being asked to manage for the first time — with no onboarding of their own. Nobody handed them a guide to directing fast, capable workers who don't ask clarifying questions unless you teach them to. Nobody explained that you have to know something about a task to know if it was done right. That's true for managing people. It's equally true for managing AI.
There's another wrinkle worth naming. My AI team has an unreliable memory. Some context carries forward. Much doesn't. And I rarely know which is which. The project history, the decisions already made, the things we tried that didn't work. I can't assume they remember. I try to build systems that create memory — structured notes, project summaries, markdown reference files, careful setup at the start of each session. Sometimes that works. Not always. Which means the institutional knowledge lives primarily with me. That's not a skill gap on their part. It's a structural reality that changes what managing them requires.
I have to carry what they can't.
The skill you cannot afford to lose
There is research emerging on what happens when people hand over too much. A 2025 study by Michael Gerlich at SBS Swiss Business School surveyed 666 people across age groups and found a significant negative correlation between frequent AI use and critical thinking ability. The culprit: cognitive offloading — the tendency to hand your thinking to an efficient tool and exercise your own judgment less. The effect was most pronounced in younger users. Which means the professionals who grew up with AI may be the least equipped to catch it when it's wrong.
Good managers don't just accept their team's first answer. They ask for the reasoning. They push back. They say "go back and prepare" when the thinking is shallow. They make their people — and themselves — sharper by asking for thorough explanation and reasoning, not just output.
The same discipline applies here. If you stop asking why, if you start accepting polished as a proxy for correct, you don't just get bad output. You get worse at recognizing bad output. The very speed and quality that makes your AI team valuable becomes the thing that quietly erodes your judgment.
The more talented the team, the more you have to lead.
Not less.
What this means if you're stuck
If AI hasn't gotten off the ground for you yet, I'd start with two questions before assuming the problem is the technology.
Are you treating AI as a tool instead of a team? And have you onboarded that team?
Onboarding should look the same as it would with any talented new hire who doesn't know your work, your standards, or what a good outcome looks like for you specifically.
And, like any good managerial relationship, have you defined how you'd like to work together? I ask my team to work with me more than for me. That distinction changes everything about what comes back.
Explain the job. Review the work before it goes out. Stay sharp enough to catch the wrong amazing when it arrives. And remember that you are now the manager of a team that is faster than you, skilled in ways you aren't, and entirely unaffected by whether any of this goes well.
That last part is yours alone.
Wendy Lynch, PhD · Lynch Consulting Ltd.
