If AI does the junior work, how does anyone become senior?

By Bryce Stuckenschneider | Published 2026-10-14 | Strategy | 6 min read

As AI takes on the work people once learned from, leaders need to design new ways for their teams to build judgment, practice, and grow.

We’re at MAICON, the Marketing AI Conference, in Cleveland this week. It brings together marketing and business leaders to work through how AI is changing their organizations.

We send our team because getting smarter is part of our job. We want to learn from people doing interesting work, build relationships with other operators building and scaling businesses, and challenge what we think we know.

Working across clients gives us a useful view of how AI is being adopted. It also gives us assumptions that need to be tested. Are other teams running into the same problems? Has someone found a better approach? What are we missing?

I want us bringing those questions into rooms like this, then bringing better thinking back to our clients.

Paul Roetzer gave us plenty to think about in his keynote, “The Architect, the Orchestrator, and the Apprentice.”

He told a story about his first job in marketing. He was writing press releases, researching, building proposals, and trying to hand something to his boss that wouldn’t come back with red marks all over it.

He learned through trial and error, with a mentor who gave him time and feedback. The work was how he got better.

“The process was the point of the work,” Paul said. And it reminded us of some of our first jobs.

It’s something we need to spend more time on as business leaders. A lot of the work we’re excited to hand over to AI is the same work people have historically done to learn their profession.

Researching a market teaches you how to evaluate information. Writing a proposal forces you to understand what a customer needs. Having someone pick apart your first draft helps you recognize the difference between something that sounds good and something that holds up.

Those assignments produced useful work for the company. They also helped develop the person doing them.

As AI takes on more of that work, we need to figure out how people will build that judgment. The SmarterX team has a write-up of Paul’s framework that explores how architects, orchestrators, and apprentices fit into that changing workplace.

At Trailblaze, we describe our approach as AI-forward, human-first. This is one of the places where that belief gets practical. It affects how you assign work, what you teach, how you review an answer, and what opportunities someone gets next.

We’re excited about the capacity AI can create. Think about the project that keeps getting pushed because everyone is buried in the week’s deliverables. Or the customer follow-up that deserves more attention than the team has time to give it. Being able to take on that work matters.

We want leaders asking what their teams could accomplish with more room to work. And we think some of that room needs to go toward helping people become more capable. Otherwise, we could get very efficient at producing work while getting worse at developing the people responsible for it.

Take someone learning to write a proposal… AI can help them get a first draft together. The learning depends on what happens around that draft.

Does the person understand the customer’s problem? Can they explain why this recommendation makes sense? Did they catch the assumption that needs to be checked before anyone makes a promise? Can they tell you what they changed and why?

A manager can use that draft as the starting point for a useful conversation. Walk through the recommendation together. Ask the person to defend it. Show them where the reasoning breaks down. Let them revise it and own the next version.

That’s a different assignment from spending an afternoon staring at a blank document. It still asks the person to think. We should be designing more of those assignments.

We can also use AI to help people practice. Give someone a customer scenario, let them work through a response, and have them explain their decisions. Use it to explore alternatives before bringing a recommendation to a manager. Let someone attempt work that would previously have been beyond their experience, with an experienced person close enough to help.

The important part is making the learning intentional. A polished output doesn’t tell you how much the person understood.

This applies to experienced people, too. When you delegate more work to AI, you have to get clearer about the outcome, the context, and the standard you’re willing to accept. Someone still needs to recognize a bad assumption, decide when an answer needs more work, and take responsibility for what reaches the customer.

Being good at that takes practice. So does teaching someone else to do it.

We think managers will need to spend more time making their judgment visible. Explain why you rejected an answer. Talk through the tradeoff. Show the newer person what you noticed that they missed. A lot of expertise lives in those small decisions. If the team only sees the finished product, they lose the chance to learn how you got there.

This connects to something we wrote recently about having an AI charter alongside your AI policy. If a company says it will help people build AI skills, that should show up in the workweek. People need time to practice, access to appropriate tools, feedback, and a chance to take on more responsibility as they improve.

If we say we’re keeping people at the center of this, we should be able to explain what that means for someone early in their career. Or someone whose role is changing after twenty years of doing it well.

They deserve more than “go learn AI.”

Paul’s story made the stakes feel very concrete. His early assignments gave him a place to start. His mentor helped him turn those assignments into experience. Over time, he developed the judgment to do bigger things.

We need to build paths like that for the way work is changing now.

Pick one role on your team. Look at the tasks AI is beginning to handle. What did someone used to learn by doing those tasks? Where will they learn it now? Who will help them?

I’d put those questions in your next leadership meeting. The answers belong in your AI plan.

And if this has you curious, spend some time with Paul and Mike Kaput’s The Artificial Intelligence Show. Paul previewed the thinking behind this keynote in Episode 215, which gives you a place to dig further into the framework.

• Listen on Apple Podcasts or Spotify.
• Watch on The Artificial Intelligence Show’s YouTube channel.

Find an episode that connects to something your team is working through. Send it to a colleague, compare notes, and bring a question into your next conversation. That’s part of how we get smarter, too.