The Executive's Guide to Running Your First AI Workshop with Your Team

The Executive's Guide to Running Your First AI Workshop with Your Team

The most effective AI workshops I've facilitated with senior leadership teams share a specific quality: by the end, the participants have done something with AI that surprised them — something they didn't expect to work, or that worked better than they thought possible, or that revealed a limitation they hadn't anticipated.

That surprise is the learning. And it's available to any leadership team willing to actually use AI during the session rather than listening to presentations about it.

This guide is for senior leaders who want to run their own AI learning session with their team — not as an AI expert, but as a curious leader creating the conditions for shared discovery.

What You're Trying to Achieve

Before designing anything, be clear about what you're actually going for. The most effective executive AI workshops accomplish three things:

Shared experience. Your leadership team has very different relationships with AI — from enthusiastic early adopters to skeptical non-users. A shared hands-on experience creates common ground and common vocabulary that makes every subsequent AI conversation more productive.

Visible learning from the top. When you, as the senior leader, use AI in front of your team — including when it doesn't work perfectly — you send a signal more powerful than any communication you could craft. You are demonstrating that learning in public is expected and that the technology is something to engage with, not manage from a distance.

Insight about what's possible. Well-designed AI sessions reliably produce moments where participants see something that changes their mental model of what AI can do for their specific context. These moments of genuine discovery are the seeds of meaningful adoption.

The Format That Works

A well-designed executive AI session has four phases.

Phase 1 — Frame (10-15 minutes). Set the context, not the agenda. Explain why you're doing this now, what you're hoping the group learns together, and what the norms are: this is exploration, not evaluation. There are no performance expectations. Everyone is a beginner here.

One effective framing: "We're here because I want us to have a shared experience with these tools before we make any more decisions about them. The goal isn't to leave with a strategy. It's to leave with better questions."

Phase 2 — Demo and orient (15-20 minutes). Show the tool briefly — not a polished presentation, but a genuine use. Pick a real problem from your current work and engage with the AI in real time. Talk through your reasoning as you go. Show how you prompt, how you evaluate the output, how you push back when something doesn't seem right.

The less polished this is, the better. Authenticity is what creates permission.

Phase 3 — Hands-on work (30-45 minutes). This is the core of the session. Give the group a real challenge — one that's relevant to their actual work — and ask them to engage with it using AI. Structure this with clear time boxes and specific deliverables:

• First 15 minutes: individually or in pairs, use AI to generate options or analysis on the problem
• Next 15 minutes: share what you got, compare results, discuss what the AI got right and wrong
• Final 15 minutes: if time allows, use what you've learned to run a second, better attempt

The problem you choose matters. Generic case studies produce generic engagement. Real organizational challenges — ones where participants have genuine stakes and contextual knowledge — produce the observations that shift mental models.

Phase 4 — Debrief and commit (15-20 minutes). Close with three questions:

1. What surprised you?
2. What are you going to try differently in your work this week?
3. What do you want to learn more about?

The answers to these questions tell you more about your organization's AI readiness than any survey. And the commitments made in front of peers have a remarkably high follow-through rate.

What to Avoid

Avoid expert presentations. The moment the format shifts to "here is what AI can do" rather than "here is us doing it," the learning potential drops by half. Passive learning about AI does not produce behavioral change.

Avoid evaluating performance. If participants feel like their AI use is being assessed — their prompts judged, their outputs compared — they will optimize for looking competent rather than learning. Establish explicitly that this is exploration, not evaluation.

Avoid overclaiming. Resist the temptation to present AI as unambiguously transformative. If participants encounter real limitations during the session — and they will — and those limitations aren't acknowledged as normal and expected, the implicit message is either that the technology is being oversold or that they're using it wrong.

Avoid using a vendor. AI vendor-led sessions have an inherent conflict of interest: the vendor has an incentive to make AI look better than it is in your specific context. Run the session yourself or hire a facilitator with no stake in which tool you adopt.

A Note on Facilitation

You don't need to be an AI expert to facilitate this session effectively. You need to be genuinely curious, willing to not know things in public, and skilled at creating conditions where others feel safe to learn alongside you.

If you're uncertain about the tools, say so. If something doesn't work the way you expected during the session, treat it as a learning moment rather than a problem. If participants come up with better approaches than you did, celebrate that explicitly.

Your job is not to teach AI. It's to lead learning. Those are very different things — and the second one you've been doing for your entire career.