Much of a company's tacit knowledge is trapped in meetings. And meetings often dumb down our collective problem-solving. But meetings are getting reinvented through AI.

Today’s default meeting is weirdly low-tech compared to what they could be:
  •  We stare at slides someone built by hand.
  •  Notes are partial, biased, and inconsistent.
  •  The loudest voice (often the most senior) can steer the outcome.
  •  “Next steps” are fuzzy, and memory becomes politics.

Intelligent AI use can flip the stack.

Now you can ask machines to:
  •  Summarize what was actually said (not what we wish was said).
  •  Critique the discussion through specific lenses (risks, assumptions, customer value, feasibility, ethics, incentives) regardless of who spoke.
  •  Detect gaps, contradictions, and unresolved decisions.
  •  Produce clean outputs: action plans, decision logs, follow-ups, even slides/infographics—without a week of manual work.

That changes the human job for the better.

In AI-native meetings, humans need to:
  •  Direct the machine like a high-performing colleague: ask better questions, request specific frameworks, push for closure.
  •  Do change management: norms, adoption, psychological safety, and accountability.
  •  Build “meeting context” intentionally: the right background docs, goals, constraints, and evaluation criteria so the model isn’t guessing.

The new meeting will be the equivalent of a factory of decisions and problem-solving, compared to the old practices, which will appear comparatively artisanal and prone to defects.

This won't happen automagically. Management of that factory requires context engineering, facilitation, and decision discipline. These are all professional skills that humans can get better at.