May 20, 2026
François Bossière, Ph.D, Founder of Polynom
Why the same AI tools that doubled developer output have barely moved the needle for executives — and what to do about it.
Same companies. Same employees. Same AI tools deployed widely across the organization. Yet software engineers ship measurably faster every quarter, while managers report saving only a few minutes a day on email. The standard explanations — change management, model maturity, training gaps — miss the structural point.
The Unit of Work Matters
For a software engineer, the unit of work is the function, the file, the pull request. AI met the engineer exactly at that unit, and that is why the gains showed up so quickly and so clearly.
For a manager, the unit of work is the meeting. Senior executives spend roughly 23 hours a week in meetings on average. The document — the email, the deck, the report — is just residue. And that is exactly where office AI has been pointed.
In other words: the productivity gap is not a maturity gap. It is a targeting gap.
Three Moves to Close It
AI inside the meeting itself. Real-time retrieval, drafting, and action capture during the conversation. Roughly 80% of the questions raised in a meeting are already answered in documents on hand — they need to be surfaced when they are needed, not afterwards.
AI on parallel work while the meeting runs. Emails answered, reports built, the next briefing prepared — so the executive hour stops being single-threaded.
An AI proxy that attends on your behalf. On a two-to-three-year horizon, an agent that sits in meetings for you and pings you only when a decision is actually required.
A Prerequisite Nobody Wants to Hear
None of this works without a real agenda. Most meetings don't have one. Fix that first — everything else builds on it.
Read the Full Argument
The meeting is the work. The document is residue. We unpack what that means for the operating model of a modern executive, the tools that matter, and what an AI-augmented executive hour actually looks like in practice.