AI and the Other productivity disaster: Meetings!

AI and the Other productivity disaster: Meetings!

We know that Artificial Intelligence can help with the worst of our productivity problems: Email Overload. After all, this is what Knowmail does: apply AI technology to go through our Inbox and find the golden nuggets – the emails we really want and need to read, in the context of whatever matters to us at a given moment. This can save us much time and frustration, indeed.

But there is another productivity disaster in organizations besides email, and it is costing them millions and reducing their effectiveness every day. This is the state of the meetings that are a central part of knowledge workers’ job. Managers, engineers and white-collar workers attend meetings every day; many of them spend most of their time in meetings, many of them are double-booked, and I’ve seen companies where people’s calendars are even triple-booked much of the time (really!). And even when they make the meetings, these are plagued by a range of ineffective practices: the wrong people are there, the right people are not, the discussions are unfocused, the decision process is unclear, the agenda is undefined, there are no minutes, and so on. Small wonder that when Ignites magazine surveyed readers a few years ago about what is the biggest waste of time in their day, unnecessary meetings got first place at 38% (with too many distracting e-mails coming in second at 25%).

So – can Artificial Intelligence come to our rescue (again)?



Not yet, not today, but I see no reason why not soon. When I advise companies on meeting effectiveness I recommend they implement various effective practices. These can be things like making sure the right people, and only they, are invited; setting the right duration/frequency of a meeting to match its content; issuing a meeting invitation X days in advance and including in the invitation an agenda and the required preparation; keeping the meeting running on its published timeline; and publishing the minutes in a timely manner.

Now, educating people to do all this is fine, but doing it takes time and effort – more than not doing so; and people in today’s hectic workplace are always busy, distracted, and at times thoughtless or just lazy. It is easy to forget or neglect to stick to all these guidelines. But a computer is never distracted nor lazy. Why not let an AI engine take care of these things? I don’t mean using a template in Outlook to ensure meeting requests have the right fields filled in. I mean using an AI program to use its knowledge of the organization’s structure and practices, the roles and interactions of various people, and experience, to really facilitate a better meeting culture.

For example:

  • Allow the meeting organizer to tell the AI tool what a meeting is intended to achieve, and let the tool decide how to set the agenda and who should be invited. How would it know? Based on knowing the organization’s approval processes (which determine what level of managers need to be present), its stakeholder map, and the mapping of roles to individual employees. Moreover, the AI would know who, among the possible reps of a department, is the best one to invite – based on their experience in the specific subject, and on their likelihood to show up given their other workload at the time.
  • Have the AI decide how long a meeting should be, given its proposed agenda and its knowledge of the specific participants’ talkativeness.
  • Let the AI define the preparation work needed for a meeting by each invitee, based on an understanding of the agenda and of the likely course of the discussion.
  • Have the AI follow on the meeting’s progress, by listening in and interpreting progress along the agenda, and have it nudge things along if the meetings gets rat-holed (a common occurrence).
  • Let the AI generate the minutes of the meeting by sitting in on it, integrating voice and presentation content.
  • And who knows what else?


Absurd, you may think: how would Artificial Intelligence know to do all this? How would it know that it is a good idea to invite Alice and not Jack to represent HR in the building expansion review meeting? Simple: this AI would have access to everything in the organization – everyone’s emails, calendars, skills, effectiveness or delinquency in filling the expected norms in the meetings they had chaired in the past, correlations between who attended what meetings and what happened there… lots and lots of data, to be sliced and diced and analyzed and understood. All automatically, with no human involvement. So if in the past all meetings that involved Jack and the head of building maintenance and lasted more than 30 minutes tended to result in raised voices – especially Jack’s – and no decisions being reached, then Alice may really be a better person to send to that meeting.

And if you think a computer would not know what a raised voice sounds like, try yelling at a voice response system next time it puts you on hold. You may be in for a surprise. The computer, like the wise owl in the nursery rhyme, may not speak much, but it hears and sees and – soon, anyway – understands everything around it. Might as well let it help us make those meetings more effective – and more pleasant!


Nathan Zeldes
Nathan Zeldes

Nathan Zeldes is a globally recognized thought leader in the search for improved knowledge worker productivity. After a 26 year career as a manager and principal engineer at Intel Corporation, he now helps organizations solve core problems at the intersection of information technology and human behavior.