Most AI workshops fail in the same way. A room full of people watches someone demo a tool, everyone nods, and a month later nobody's using it. The demo was impressive. It just wasn't about their work.

I build workshops the other way round. We start with the tasks on your team's desks this week and finish with each person running a workflow they'll use again. Here's where most teams are starting from, how I run a session, and what your people leave with.

Where your team is probably starting from

Pew Research Center asked US workers how often they use AI chatbots such as ChatGPT, Gemini or Copilot at work. About one in ten said every day or a few times a week. Another 7% said a few times a month. And 55% said they rarely or never use them.

Age makes a difference. Among workers aged 18 to 29, 23% use chatbots at least a few times a month, compared with 17% or less in older groups. And among people who don't use AI yet, 31% said at least some of their work could be done with it.

So in a typical room you'll have a few keen users, a lot of people who've tried it once or never, and a group who suspect it could help but don't know where to start. A good workshop serves all three without boring the first group or losing the last.

What people use AI for at work

The same Pew survey asked chatbot users what they use them for. The most common answers were research, at 57%, editing written content, at 52%, and drafting written content, at 47%. Those three jobs are the backbone of almost every workshop I run.

Pew found something else I build every session around. Four in ten chatbot users, 40%, said the tools were extremely or very helpful for doing things more quickly. Only 29% said the same about improving the quality of their work.

That tells you what to teach. Speed is the easy win, so I show people how to get it. Quality is where they need help, so we spend real time on checking, editing and writing better briefs, so the output is something you'd put your name to.

• Research: 57% of chatbot users.
• Editing written content: 52%.
• Drafting written content: 47%.

How I run an AI workshop

Every session follows the same arc, whatever the team or format.

1. Before the day: I ask each attendee for two tasks that eat their week, so the examples are theirs.
2. Start with one job: we take a common task and do it live with AI, step by step.
3. Hands on: everyone runs the same workflow on their own work, with me in the room to help.
4. Check the output: we compare drafts, spot what the AI got wrong and fix the brief.
5. Leave with a plan: each person writes down the one workflow they'll use that week.

Formats that suit different teams

A single half day works for teams who need a confident start. A full day suits teams ready to build several workflows. A short series, with a session every week or two, suits teams who learn best by trying things between sessions and bringing questions back.

I run workshops in person and over video, always in English. For larger organisations rolling AI out across departments, my page on corporate AI training explains how I plan training at scale.

Mistakes I see in AI workshops

The biggest one is teaching tools instead of tasks. A tour of every feature in a chatbot feels thorough, but people forget it by lunch. One task done well, on their own work, is what they remember.

The second is skipping the rules. Teams need to know what they can and can't paste into an AI tool, especially client data, before they start. The third is treating the workshop as the finish line. Without a follow-up, most people drift back to the old way within a few weeks, and nobody notices until the budget review.

What your team leaves with

I judge a workshop by what happens the following month, not by the feedback forms on the day.

• A working prompt library: briefs for the tasks your team does most, written during the session.
• One workflow each: something every attendee has already run on their own work.
• A checking habit: a short list of what to verify before anything AI-made goes out.
• A clear rule set: what the team can and can't put into AI tools.

My advice before you book AI workshops

Ask whoever runs your workshop to use your team's own tasks, not generic demos, and to send attendees a short question beforehand. Then plan a check-in four weeks later. If people can't name a workflow they're still using by then, the workshop didn't work, however good it felt on the day. My guide on how to tell if AI training worked shows what to measure.

To talk through a workshop for your team, see my work with me page.

Frequently asked questions

What is an AI workshop?

A hands-on session where a team learns to use AI tools on its own work. Lilach Bullock's AI workshops start with each attendee's real tasks and end with a workflow they'll keep using.

What do workers use AI chatbots for?

In Pew Research Center's survey, the most common uses were research (57%), editing written content (52%) and drafting written content (47%).

How long should an AI workshop be?

A half day suits a confident start, a full day suits building several workflows, and a short series suits teams who learn by trying things between sessions.

Can AI workshops run online?

Yes. Lilach Bullock runs AI workshops in person and over video, in English.

How do I know if an AI workshop worked?

Check four weeks later whether people are still using at least one workflow from the session on their own work.