Should your team do an AI workshop or an online course?

Nov 7, 2026
Harry Siggins

Once you've decided your team needs to get better at AI, the next decision is how. You could buy an online course, bring someone in to run a workshop, or do both. And leaders often ask me how a workshop is different from a course, or from training they've already done.

My answer starts with a question. There's already a huge amount of AI content out there. If it's out there, why hasn't your team picked it up already? In my experience, it's because they haven't seen it against their own work, or used it in a way that makes it theirs. That's what a workshop built around your team's real work can do. To see why, start with what happens when a team tries general material on its own work.

Why more tools and content don't change how people work

Most AI courses teach general skills on general examples, which is useful up to a point. But when a team tries it on their own work, they reach the places where their company does things its own way, and they get stuck. They could probably work it out on their own. But they have a day job, and what they need at that point is a bit of help getting through it.

Giving people AI tools doesn't, on its own, change how they work either. In a field experiment by researchers from Microsoft and Harvard Business School, across 66 firms and more than 7,000 knowledge workers, people were given an AI tool built into the email, meeting and writing apps they already used. Those who used it spent about two fewer hours a week on email. But beyond those individual time savings, the researchers didn't detect any shift in how much work people did or what kind of work it was. One explanation they offer is that email is something each person can change alone, while changing how meetings or shared documents work means colleagues agreeing on new ways of working together. That's the part a team has to work out together, and a workshop on their own work gives them the room to do it.

None of this makes courses useless, though.

What a course is good for

Courses have a place. A good self-paced course can give people some tactical depth on particular tools, and a free course from an AI company is a low-cost way to learn the basics. The catch is that this field changes quickly, so course material goes out of date fast, which is why I hesitate to lean on courses.

If a course is the do-it-yourself end, the other end is having someone build it for you.

What about someone who builds it for you?

A consultant who builds AI tools for your team will usually get you a result faster. The trade-off is that your team doesn't learn to do it themselves, and what's built needs keeping up to date as the tools change every few months. That's why I don't build for clients. I'd rather the team can carry on without me, which means the workshop itself has to change how they work.

What a workshop should change

Leaders often ask whether a workshop is just training by another name. To me, they aim at different things. Training takes the steps that worked for someone successful and copies them for everyone else, which builds a habit but on its own doesn't change how people do their own work. A workshop should change exactly that. I call that kind of change enablement, and it has three parts.

Raising the floor, so everyone has the same basic understanding of what AI is and what it can do.
Raising the ceiling, which means giving people room to experiment on their own work and see what works and what doesn't.
Leadership. Leaders who understand the opportunity, and take part, keep the first two from slipping.

Here's what that looks like in practice.

What a good workshop looks like

I'll use my own as an example. Before anything starts, I have a short call with each person, about 15 minutes, to understand their work. Then there are three two-hour sessions, a week apart, with a smaller group. The group stays small because hands-on work stops being hands-on when there are too many people in the room. It turns into a lecture.

Across the three sessions, the work builds like this.

In the first session, we build a shared foundation, a common language for talking about AI and a sense of what today's tools can do. The group usually works hands-on with a practice set of data I build to look like their company's work.
Before the second session, everyone names the biggest problems they're trying to tackle. In the second, I work through how I'd approach some of those problems, live, in front of the group.
Between the second and third sessions, people build, experiment, succeed, and fail, and the third session goes further, into more advanced work.

The week between sessions is one of the most important parts of the design, because it's when people take what they've seen back to their own desk and try it on their own terms.

People often ask how technical their team needs to be. Being good with AI now has much less to do with writing code than with understanding what AI agents can do, how to direct them, and how to judge what they produce. Nobody in the room needs to code.

Kelly Scheib, Chief People Officer at Crunchbase, described the change after her team's workshop. "AI has not replaced my judgment, but it has made the first pass dramatically faster and helped me see patterns, ask better questions, and spend more time on the human and strategic parts of the work."

If that's the kind of change you want, the next question is who should run it.

What to look for in whoever runs it

These are the five things I'd look for. I run these workshops myself, so weigh the list with that in mind.

They've worked across many functions, so they can talk to every team in one common language. If you bring in a different specialist for each team, each team learns its own vocabulary and its own way of using AI, and the teams can't easily share what they learn.
They can think critically and quickly about your team's actual problems, so what they teach is specific to your team's work.
They run both company-wide and team-specific workshops, so they can go broad or deep depending on what you're trying to achieve.
They're dedicated to keeping up with how AI is changing, beyond engineering, and they can tell you what it means for the way your team works.
They build trust with your team, understand their problems, and are someone your team can turn to during and after the sessions.

How to decide

If your team has already had access to AI tools, courses and content and still hasn't changed how they work, more of the same is unlikely to change that. Look for a workshop built on their own work, and use the five points above to choose who runs it.