Something subtle has been happening in the organizations I work with over the past eighteen months. Meetings are getting shorter, but the preparation time before meetings is getting longer. The ratio has flipped from what most meeting productivity advice assumes.
Before AI tools, the typical meeting dynamic was: people showed up with partial information, spent the first fifteen minutes getting everyone on the same page, then spent the next thirty actually making decisions. The meeting was where context got shared.
Now the teams that are using AI well are showing up to meetings already contextualized. The AI has summarized the relevant documents, pulled the key data points, identified the open questions from the last meeting, and surfaced conflicts between what different team members said they were working on. The meeting starts with shared context already established.
This sounds like a straightforward productivity improvement. And in terms of decision quality and meeting length, it is. But it has created a new problem that nobody warned anyone about.
The people who are not using AI tools are getting left behind in real time.
They show up to meetings where everyone else has done AI-assisted preparation and they are the ones trying to catch up on context while others are already three steps into the decision. The preparation asymmetry has become a participation asymmetry.
I have watched this happen in a leadership team I was advising. One executive had fully adopted AI for meeting preparation and was consistently the most informed person in the room. Two others were still doing manual preparation. The dynamic was not hostile, but it was visible. The AI-prepared executive was operating at a different altitude than the others, and the others knew it.
The thing that resolved it was not mandating AI tool adoption, which creates its own problems. It was shifting where the AI-generated preparation work landed. Instead of each person doing AI prep individually and arriving with individually curated context, the team started sharing AI-generated pre-reads before every meeting so that the context was collective rather than individual.
That shift sounds small. The effect was significant. It moved AI from being a competitive advantage within the team to being a shared infrastructure for the team. The meetings got better for everyone instead of better for some people at the expense of others.
The broader lesson I took from this is that AI tool adoption inside a team is not a personal productivity question. It is a team dynamics question. The tools change how people relate to information, and when different people on the same team relate to information differently, the team dynamics shift in ways that create friction nobody anticipated.
The organizations that are handling this well are treating AI adoption as a team-level intervention rather than an individual-level one. They think about what the team's shared information environment looks like rather than optimizing each person's individual information environment. They make AI-generated context a shared resource rather than a private advantage.
This requires explicit design. It does not happen by default.
Top comments (0)