The Workshop That Starts With Your Worst Task

Here's how most AI training workshops begin: a facilitator opens a slide deck, introduces the tool, walks through its features, and shows a demo. Maybe there's a hands-on exercise where everyone builds the same thing. People leave with a certificate and a vague intention to try it on Monday.

Here's how the best one I ever ran began: "What's the task you dread most every week?"

Twenty people in the room. Twenty different answers. Someone spends four hours every Monday reconciling data between two spreadsheets. Someone else manually writes the same status update email to three different stakeholders every Friday, changing six words each time. Someone reformats the same report template every month because the source data always arrives in a slightly different layout.

Those answers became the workshop.

Features don't stick. Pain does.

There's a reason feature-led training has such a dismal track record for sustained adoption. When you teach someone that an AI tool can "summarise documents, draft emails, analyse data, and generate creative content," you've given them a menu. Menus are overwhelming when you're not hungry.

But when someone walks in already frustrated by a specific task — when they can feel the two hours they lose every week — and you show them how to eliminate that exact problem? That's not a feature demo. That's relief. And relief is memorable.

The difference is who does the translation work. In a feature-led session, the learner has to map abstract capabilities to their own context. Most won't bother — not because they're lazy, but because it's genuinely hard to see how a generic demo applies to your specific Tuesday afternoon. In a pain-led session, you've already done that mapping. The tool arrives as the answer to a question they were already asking.

How it works in practice

Before the session, you survey participants. Not "what do you know about AI?" — that just creates anxiety and posturing. Instead: what are you spending time on that feels repetitive? What task would you love to never do again? Where do you copy information from one place to another?

You cluster the responses. Patterns emerge quickly. There are usually three or four common categories: data reformatting, repetitive communications, manual reporting, and information lookup across systems.

The workshop then follows a simple structure. You take real examples from the room — anonymised if needed but recognisable. You work through them live: here's the pain, here's an approach, let's build the first version together, let's test it, let's see what breaks.

Nobody needs to be told why this is relevant. They can see their own problem being solved in front of them.

The learner who already knows everything

Every workshop has someone who considers themselves advanced. In a feature-led session, they're bored — they already know the interface, they've already tried the basic capabilities.

In a pain-led session, the difficulty scales naturally. The basic user learns that AI can help with their weekly report. The advanced user discovers that their "solved" workflow still has three manual steps they never questioned. The conversation shifts from "can the tool do this?" to "should this process exist at all?" — which is a question that challenges everyone regardless of experience level.

The output that matters

At the end of a feature-led workshop, people have notes about capabilities they might try someday. At the end of a pain-led workshop, people have a working solution to a problem they walked in with. They don't need motivation to use it on Monday. They're already planning to.

The best training doesn't teach tools. It eliminates problems. Start with the worst task in the room, and the tool becomes the obvious answer rather than an abstract possibility.