By lunchtime, the executive committee was reading its own thinking on the screen.

Senior workshops usually fail in one of two ways. A pre-workshop survey leaves the committee debating yesterday's answers in today's room. A round-the-table share gives the airtime to the loudest voices and synthesises nothing at the end. Neither produces something the committee can act on the same day. The firm wanted decisions, not slides.

During the morning, an adaptive AI agent interviewed each committee member in parallel, with one focused question per AI readiness dimension and follow-ups wherever an answer signalled real intent or real friction. By mid-day the synthesis read back, dimension by dimension: where the committee agreed, where it split, and which minority ideas deserved executive attention.

The top lines were no surprise. The data agenda dominated, leadership tone mattered, capacity was a constraint. What was useful was seeing the exact share of the committee holding each position, and where the real dissent sat. Three minority ideas carried more energy than headcount: an AI lab outside current constraints, digital twins, and a long-horizon AI ambition. None would have survived a one-minute round-the-table share. The synthesis surfaced each one with its proponents, instead of letting the average bury it.

The afternoon moved from describing positions to deciding actions, with owners assigned. The committee's own thinking had become the agenda. No other shape of workshop the firm had tried had managed that.