Each row is a claim that several studies support together, and each column is one study. A filled dot marks a study that is a main source of evidence for the claim. An open dot marks one that adds supporting evidence, and a diamond one that qualifies or complicates it. Pick a claim to read what the studies show, how strong the evidence is and what it does not show. Pick a study to see every claim it feeds.
Meeting purpose is usually implicit, and its stand-ins hide it
Across interviews, recorded meetings and a field experiment, people rarely said what a meeting was for. Topic lists and standing agendas stood in for goals, saying what would be discussed but not why.
Evidence: Interviews, recordings of real meetings and open-text survey answers, all at one global technology company.
Does not show: How common unstated purpose is outside one technology company, or in other kinds of organisation.
● Core evidence
- Mental Models Many people treated a topic list or a recurring meeting's standing agenda as its goal, which hides why the meeting exists.
- Are We On Track Only 4 of 12 recorded team meetings stated what they should achieve, though most had an agenda.
○ Supports
- What Does Success A recurring meeting rarely had specific goals for any one instance, so reflection on it felt too general.
- Nudging Lack of time before meetings was the most common barrier to thinking about goals, for organisers and attendees alike.
Purpose is held unevenly, so intentionality is a social problem, not an individual one
Who may ask why, who notices drift and who can change a meeting all differ by role. Organisers and attendees wait on each other, so support has to be reciprocal and safe, not aimed only at organisers.
Evidence: Interviews and technology probes, with open-text answers from a field experiment, at one company.
Does not show: Whether changing roles or norms would make people more willing to ask why or to flag drift.
● Core evidence
- Mental Models Invitees avoided asking organisers what a meeting was for, for fear of seeming impolite, critical or ignorant.
- Are We On Track Junior staff and people not running the meeting noticed drift but felt it was not their place to say so, and feared being flagged publicly.
- Nudging Attendees felt unable to change meetings they had not organised, and organisers assumed attendees would not read goals. Each waited on the other.
○ Supports
- What Does Success Attendees often felt they could not act on any solution in a meeting someone else organised.
- CoExplorer probe People wanted AI changes to the plan to go through the organiser, and some worried dominant speakers could steer the meeting.
Putting purpose into words is the active ingredient, and repetition seems to train it
Whenever people stated or checked a meeting's purpose, they reported clearer goals and better preparation. The one field experiment found no gain in rated effectiveness, so the benefit rests on people's own accounts.
Evidence: Technology probes and a preregistered field experiment with 361 employees and 7,196 meetings, with self-reported outcomes.
Does not show: That articulating purpose makes meetings more effective. The field experiment found no gain in rated effectiveness, and there was no group without any survey.
● Core evidence
- What Does Success Putting implicit goals into words helped people focus, warm up and arrive with a plan, and some then shared summaries or agendas.
- Nudging Goals got clearer and better communicated over two weeks of rating meetings, as if repetition were micro-training. An extra pre-meeting prompt added no measurable gain in rated effectiveness.
○ Supports
- Are We On Track Watching their own meetings with AI support, most people first asked themselves what the meeting was for.
- Mental Models Supporting intentionality is about reflection, not forcing explicit goals onto every meeting.
Timing and dose decide whether support helps or hurts
Support helped when it came early, matched the stakes and stayed in the background until needed. Prompts that came too late, too often or too loudly were ignored or resented, and some meetings work better without fixed goals.
Evidence: The strongest convergence: interviews, three technology probes and a field experiment point the same way.
Does not show: The best timing or frequency for prompts. Studies compared few options, and most evidence is self-reported.
● Core evidence
- Nudging Prompts five minutes before a meeting were often too late to act on, and prompting every meeting caused survey fatigue.
- What Does Success People saw important, uncertain meetings as the best candidates for reflection, and reflecting on every familiar recurring meeting as excessive.
- Are We On Track A passive display of goals was easily overlooked, while pop-up questions prompted action but could intrude.
○ Supports
- CoExplorer probe People wanted the AI to ask before showing files or apps, yet wanted fewer prompts during the meeting.
◆ Complicates
- Mental Models Goals cut both ways. Goal-light meetings allowed discovery and connection, and stated goals could feel like pressure.
AI helps most as a coach and an assistant, not as a decider
People valued AI that asked good questions, drafted structure and surfaced drift, but distrusted AI that acted on its own. Keeping people in charge, and checking the AI's output, came up in every study that built a system.
Evidence: Technology probes and design prototypes. The quality of AI output was not formally evaluated.
Does not show: How well the AI performed. Its outputs were not formally evaluated, and most probes used simulations or recordings.
● Core evidence
- What Does Success An assistant that asked questions rather than giving answers made people think for themselves, though solution-minded people wanted recommendations.
- CoExplorer probe People were sceptical of AI-generated content before it made any mistake, and wanted the organiser to keep the final say.
○ Supports
- CoExplorer 2D and 3D Generative AI could draft a meeting's aim and phases from a short invitation, but tended to propose changes when none were needed.
- Are We On Track People wanted control over goals and over AI corrections, and explanations for AI nudges.
- Temporal Work In early, unevaluated trials, an AI model found many of the links between meetings that our own analysis had found.
Purpose has to travel, across the meeting lifecycle and from one meeting to the next
A meeting's purpose is set when it is scheduled, tested during it and needed again afterwards and at the next meeting. Today's tools keep each moment separate, so goals should flow from invitation to facilitation to recap.
Evidence: Analysis of recorded meetings, interviews and design prototypes. The proposed designs have not yet been tested with users.
Does not show: That chained recaps and handoffs help teams. These designs have not yet been tested with users.
● Core evidence
- Temporal Work People looked back and ahead throughout their weekly meetings, yet existing tools covered only a narrow slice of time.
- Mental Models Tools scatter a meeting's context across scheduling, chat and meeting apps, so goals should surface from scheduling onward.
○ Supports
- Nudging People said stated goals tied their work together, and saw them as useful input for AI meeting tools.
- CoExplorer 2D and 3D AI could carry the plan into the meeting, bringing up each phase's files and apps as the talk moves on.
- CoExplorer probe Design should span the whole meeting lifecycle and the wider project, for example sharing the focus tool well in advance.