Meetings have long been the “necessary evil” of work life. But recently, they’ve crossed a line. They’re worse than ever.
Between 2019 and 2024, my colleagues and I found that the time workers spend in unproductive meetings surged, with the steepest rates now among managers. And now there’s a new layer of dysfunction and distraction: it’s more and more common to enter a meeting where many of the “attendees” aren’t people at all, but bots.
New insights—based on several years of analyzing collaboration data across customers—help explain how we got here and why managers sit at the center of both the problem and the leverage point for fixing it.
We love to blame meetings for everything that’s wrong with work. But meetings are rarely the root cause of dysfunction. They’re more often a symptom of something deeper: a broken communication system. When organizations lack clear norms for when to use Slack, email, shared documents, or asynchronous updates, meetings become the default—not because they’re effective, but because they’re the most reliable way to force alignment and command attention.
Managers feel this breakdown most acutely. They sit at the center of the organization, expected to pull information up to senior leaders and push clarity back down to their teams. When the surrounding system fails, meetings become a coping mechanism. According to Worklytics data, managers are four times more likely than individual contributors to initiate meetings. They spend roughly twice as much time in them—about 14 hours a week, nearly two full workdays—and collaborate with about 50% more people.
The cost of this meeting-first culture isn’t just calendar bloat. It’s organizational velocity. When managers become the primary clearinghouse for information, work slows down. Decisions wait on availability. Context piles up in meetings instead of moving through systems. On average, managers take about 50% longer than individual contributors to respond to internal messages from collaborators—creating bottlenecks for their teams and compounding delays across the organization.
When a communication system breaks, the problem isn’t just that organizations schedule more meetings. It’s that meetings get bigger. As Stanford Professor Emeritus Bob Sutton and University of Virginia Professor Leidy Klotz have shown, humans suffer from “addition sickness”: a bias toward solving problems by adding rather than subtracting. In organizations, that instinct doesn’t just generate more meetings—it pulls more people into them.
Managers attend about 1.5 times as many large meetings (10 or more people) as individual contributors, often with the same group, week after week. Often, these meetings aren’t designed for speed or decision-making. They’re designed for visibility and risk management—ways to ensure everyone has “been in the room” when the underlying system can’t reliably carry context, clarity, or commitment on its own.
As meetings swell, responsibility diffuses. Accountability gets murkier. What looks like collaboration is often just insurance: a roomful of people compensating for a communication system that no longer does its job.
Now add AI to the mix. AI bots and notetakers don’t cause bad meetings, but they magnify existing dysfunction, especially over-attendance and passive participation. They make it even easier to disengage (“I can read the transcript later”), turning presence into performative compliance rather than active contribution.
And this behavior cascades. According to Worklytics data, when managers use AI notetakers in meetings, their direct reports are significantly more likely to adopt them as well—scaling meeting habits, good or bad, through the organization.
AI isn’t inherently good or bad; its impact depends on what leaders design it to replace, and what they let it reinforce. Worklytics data shows a clear divergence in how companies are using AI for meetings and collaboration.
In one camp, AI is used to scale communication. Meetings remain the primary way information moves and uncertainty gets managed, but AI extends their reach. Notes, summaries, and action items are automatically generated and pushed to broader audiences via email, Slack, and shared documents. Information that once moved through a small working group now gets broadcast far beyond the room. In this pattern, heavy AI users tend to have working groups that are 1.2–1.8 times larger than non-AI users—a signal that AI is compensating for a system that keeps pulling more people in instead of clarifying who actually needs to be involved.
In the other camp, AI is used for information retrieval. Instead of asking people for updates, context, or status, employees ask the system. AI surfaces relevant documents, prior decisions, owners, and next steps in real-time without convening a meeting or widening the audience. In this pattern, heavy AI users tend to have working groups 0.6–0.8 times the size of those used by non-AI users—evidence that AI is being used to reduce coordination overhead by removing meetings from the critical path altogether.
Together, these two patterns show that the future of meetings isn’t so much being shaped by AI itself, but by the coordination design leaders choose to reinforce through it.
That’s why visibility matters. Without clear insight into how meetings actually operate (who initiates them, where participation overlaps, and where time gets stuck), leaders end up treating symptoms while the system keeps failing underneath. Analytics from platforms like Worklytics surface those hidden patterns, giving managers the chance to redesign how work moves, not just to run better meetings, but to need fewer of them in the first place.
