You're a Manager Now (But Most Don't Know It)

In a recent interview with Sam Altman, Microsoft CEO Satya Nadella described the future of work in five words: "Macro delegation and micro steering."

At Moderna, that future is already here. Brice Challamel, Head of AI Products, told us on Beyond the Prompt: "There are no more individual contributors at Moderna. Everyone is now a team of five, minimum."

Think about that for a second. You, an assistant, a coach, an expert, and a creative partner. Not someday. Right now.

Garry Tan, CEO of Y Combinator, confirms the trend: many roles that once involved manual human involvement will soon look more like supervising teams of agents. Andrej Karpathy, co-founder of OpenAI, noted that even as AI transforms jobs, many will be refactored into manager or supervisor roles rather than disappearing entirely.

Here's the uncomfortable truth most people are missing: you didn't just get new tools. You got a management job you never applied for.

The Problem Nobody's Talking About

When people say "just treat AI like your digital assistant," they're making a massive assumption: that you know what to do with an assistant.

But most individual contributors have never managed anyone. They've never had an assistant. They've never led a team. They've certainly never practiced "macro delegation and micro steering."

I've worked with thousands of professionals who are exceptional at doing the work. Put them in charge of guiding someone else to do the work? They freeze.

And here's what makes it worse—when their "assistant" doesn't nail it on the first try, they don't think "I need to steer better." They think "This assistant is useless."

If you had a human teammate who misunderstood your request, you wouldn't fire them. You'd clarify. You'd give more context. You'd iterate. You'd steer them toward the right outcome.

But with AI? One bad output and people conclude it can't do the job.

That's not an AI problem. That's a steering problem.

Welcome to Your New Role

You're a manager now. Not because you got promoted. Because the game changed.

The macro delegation part is easy—everyone's already delegating work to AI. "Write this email." "Analyze this data." "Summarize this document."

But the micro steering? That's where most people are failing.

Micro steering means staying actively involved in guiding the work. It means reviewing outputs, giving feedback, raising expectations, and iterating until you get what you need. It's not set-it-and-forget-it. It's active supervision.

And if you've never managed before, you're learning on the job—whether you realize it or not.

The good news? You already know what good managers do. You've had them (or wished you had them). The practices that make human managers great are exactly what make AI managers great.

The five practices below are what micro steering actually looks like in practice. These aren't AI tactics. They're management fundamentals. And mastering them is the difference between treating AI like a disappointing tool and treating it like a high-performing team.

What Micro Steering Actually Looks Like

1. They name their team members

Russ Somers, CMO at a tech startup, has Betty Budget, Aiden Adman, and Roger RevOps. When I asked why he names his GPTs, Russ told us: "When I'm interacting with 'Betty' versus 'Budget Analysis Tool,' I naturally provide more context, explain my needs more clearly, and engage in more of a dialogue. The results speak for themselves."

Naming personifies AI. And personification changes how you show up to the collaboration.

You wouldn't bark commands at Betty the way you'd type keywords into a search bar. You'd treat her like a person. You'd give context. You'd be clearer about what you need and why it matters.

That's not a trick. That's basic management. And it's essential for effective micro steering—you can't steer a faceless tool, but you can guide Betty toward better work.

2. They schedule time to review work

Here's what nobody tells you about macro delegation: your ability to metabolize information becomes the bottleneck, not AI's speed.

AI does in 20 minutes what used to take a team two weeks. You ready to review that output? Probably not—because you're still operating like you have days to respond.

Good managers block time to review their team's work. They don't let deliverables pile up in their inbox, and they don't let creative sparks die because "I'll get to it when I have time."

If you delegate work to AI, put 30 minutes on your calendar to actually review it. Treat the output like it came from a human teammate who's waiting for your feedback.

Because here's the thing: AI is waiting. And every day you don't review that work is a day you're not steering. You're just delegating into a void.

3. They give feedback and iterate

When a human teammate's first draft misses the mark, good managers don't throw it away. They give specific feedback: "This lacks X, needs more Y, the tone is too Z."

Then they ask for a revision.

Most people don't do this with AI. They get one mediocre response and move on. Or worse, they use the mediocre response because "it's good enough."

That's settling. And settling is the opposite of steering.

Here's what I do: when AI gives me something that's not quite right, I tell it exactly why. "This is too formal for my audience. I need more concrete examples. The structure buries the main point."

Then I ask it to regenerate based on that feedback.

You know what happens? The second version is almost always dramatically better. Not because AI suddenly got smarter, but because I steered better.

Micro steering isn't one prompt. It's a conversation. It's iteration. It's active guidance until you get what you need.

4. They raise expectations

AI performs to your expectations. Low expectations produce lazy prompts. Lazy prompts produce mediocre outputs. Then you conclude AI isn't that useful.

Russ Summers calls this the most dangerous thing about generative AI: "It will perform to one's low expectations." When AI doesn't nail something, the first question isn't "Can AI do this?" It's "Did I expect enough to guide it well?"

Good managers believe in their team's capability. They push for better work because they know their people can deliver it.

With AI, this is even more true. The ceiling is way higher than most people think. But you'll never hit it if you don't expect it to exist.

Ask yourself: are you treating AI like an intern who can't be trusted? Or like a highly capable teammate who just needs clear direction and high standards?

Micro steering means continuously raising the bar. Not accepting the first draft. Not settling for "good enough." Expecting excellence and guiding the work until it gets there.

5. They extend the benefit of the doubt

When a human teammate doesn't nail something, good managers assume they misunderstood—not that they're incompetent.

They clarify. They provide more context. They iterate.

But when AI doesn't nail something? Most people assume it can't do the job.

I watched this play out with my own dad. He asked me if AI could convert a 100-page PDF to CSV. I confidently dropped it into ChatGPT with a basic request. Got garbage—a single-line CSV that was completely useless.

I could have concluded "AI can't do this." Instead, I added one sentence: "Before you answer, walk through your thought process step-by-step."

Result? A perfect multi-hundred-line CSV with all the necessary categories.

AI hadn't suddenly gotten better. I had just steered better. Managed better.

When your teammate misunderstands, you don't fire them. You give them another shot with better guidance. That's the benefit of the doubt. And your AI teammates deserve it just as much as your human ones.

That's micro steering in action: assuming capability, providing better direction, and guiding toward success rather than concluding failure.

The Shift Nobody Sees Coming

Here's what's wild: everything I just described—naming, scheduling reviews, giving feedback, raising expectations, extending benefit of the doubt—these aren't AI tactics.

They're fundamental management practices.

Satya Nadella is right. The future of work is "macro delegation and micro steering." But most people only hear the first part. They delegate the work and walk away. They don't realize that macro delegation without micro steering is just abdication.

And if you've never managed before, you're learning right now whether you realize it or not.

Some people are learning well. Russ Summers tripled his department's effectiveness by treating his "GPTeam" like actual team members who need direction, feedback, and high expectations. Brice used his AI team to prepare for 16 parent-teacher conferences—historically dismal for him—and got more compliments on his teacher engagement than ever before because he could focus on the conversation instead of scrambling for notes.

Other people are learning poorly. They're barking commands, accepting first drafts, and wondering why AI isn't delivering.

The difference isn't the AI. It's the steering.

Macro Delegation Is Easy. Micro Steering Is Everything.

You're a manager now. Not because you chose to be, but because you have a team—an assistant, a coach, an expert, and a creative partner—waiting for your direction.

The question isn't whether you're delegating work to them. You are.

The question is: are you steering?

Are you naming your team members, or treating them like faceless tools?

Are you blocking time to review their work, or letting their output pile up unread?

Are you giving specific feedback and iterating, or settling for "good enough"?

Are you raising your expectations, or assuming they can't deliver?

Are you extending the benefit of the doubt, or giving up after one bad response?

The future Satya Nadella is describing—where work becomes "macro delegation and micro steering"—isn't coming. It's here. Every person reading this already has a team to manage.

The only question is: are you steering them well?

Related: It's Not An AI Problem. It's A You Problem.
Related: Beyond the Prompt: Brice Challamel
Related: Beyond the Prompt: Russ Somers

Join over 27,147 creators & leaders who read Methods of the Masters each week.