Lately, I keep hearing the same line at work: “I put this into AI and it came up with this model / framework / analysis.” A few years ago, those same problems would have triggered a working session: get the right people together, map what we actually know, and design options. Now the “thinking” arrives by email, pre‑packaged by a chatbot. For the people who used to plan and run those sessions, it doesn’t feel like progress. It feels like being cut out of our own jobs.

The line that replaced workshops

Here’s the pattern my peers and I keep seeing across teams: an AI-generated problem statement lands in an email and gets treated as the new truth about the situation, or a framework that “came out of AI” gets shared as a model we should adopt and implement. Everyone’s busy, under pressure, and being told this is the modern shortcut to insight. What used to be the starting point for a workshop is now treated as a finished product. The actual working level never gets to do the work.

From workshops to work‑about‑work

Before this pattern took hold, I could run more working sessions with my staff and peers. (Especially my favourite: The Unstuck Meeting) Someone would bring a messy problem; we’d pull together a group, put the mess on the table, and spend time understanding it: who is affected, where the constraints are, what we don’t know yet. We’d leave with decisions, experiments, or at least a shared map of the problem.

My peers tell me the same thing across different teams: they used to be asked to convene, to research, to frame the problem.

Now a lot of that has been replaced by what I’d call meta‑work:

• decoding AI‑generated problem statements to fit our our context.
• translating AI‑spit frameworks into something that might work in reality.
• sitting through meetings with over-packed AI-generated agendas.

There is less time to think, and more artefacts to react to: more decks, more emails, more models.

But less actual opportunity to apply expertise and experience.

What gets lost when AI replaces working sessions

What we’re losing in the process:

• Discovery: Time to understand the problem space: talking to people, looking at real artefacts, checking how things actually work today. Identifying the right problem to solve.
• Sense‑making: The messy process of putting evidence, perspectives, and constraints together in a group to build a shared picture of what’s going on.
• Option design: Building scenarios and trade-offs that fit reality, capacity, and risk.
• Ownership: Working through a problem together and owning the decisions versus executing the bot’s model.

How AI could support workshops, not replace them

The point here isn’t “never use AI.” The point isn’t “never use AI.” AI could support workshops, not replace them: as pre-work to outline angles, generate questions, or summarize prior documents; as documentation to clean up notes and format outputs after the session; as comparison to check team framing against AI alternatives. AI is one tool at the table, not the voice that replaces it.

Small moves from the working level

If you’re seeing the same pattern, you’re not imagining it. AI isn’t just speeding up your work; in a lot of places, it’s being used to skip the part where your work would have happened at all. Most of us are not going to change organizational behavior but we can make small moves in our own practice to protect space for real collaboration.