AI & Events, Part 1: How AI-able Is Your Event?

Three perspectives on using AI effectively: organisers, PCOs and MICE professionals, and attendees.

An organiser wants to understand why registrations are slowing. A project manager is reconciling three versions of a supplier brief. A delegate is trying to choose between six sessions happening at the same time.

They are involved in the same event, but they need very different kinds of help.

That is a useful starting point for discussing artificial intelligence in event management. Before choosing a tool, identify whose problem it should solve, what information it needs, and what a better result would look like.

A PCMA pulse survey of meeting professionals on generative AI describes applications ranging from content creation to planning support, alongside concerns about data security and preserving the human qualities of events. The practical question is how to turn those experiments into dependable working methods. (PCMA: How Event Planners Are Using Gen AI)

This opening article in our AI & Events series looks at the opportunity from three perspectives. The roles can overlap within one organisation; separating them helps clarify the work.

1. The organiser: making better decisions about the event

The organiser holds responsibility for the event’s purpose. What should it achieve? Who should attend? What would make it worth repeating?

AI can help organise the information behind those decisions. Possible starting points include grouping previous feedback by theme, comparing a proposed programme with the event’s objectives, and drafting different messages from an approved event brief.

Consider a congress planning its next edition. Its feedback contains comments about session depth, repeated topics, insufficient discussion time, and difficulties moving between rooms.

An AI-assisted review could group these comments, identify passages supporting each theme, and prepare questions for the programme committee. The committee would then examine the original responses, check whether the patterns are representative, and decide what to change.

That last step matters. Ten similar comments do not necessarily represent ten different people, and a frequently mentioned inconvenience may be less consequential than a serious issue raised once.

A useful first application: prepare a traceable summary of participant feedback for the planning meeting.

What to measure: the time required to produce a checked summary, the corrections it needs, and whether it helps the team make specific improvements.

The organiser’s responsibility remains to decide which outcomes matter and how competing needs should be balanced.

2. PCOs and MICE professionals: keeping delivery coordinated

For professional congress organisers and teams working across meetings, incentives, conferences, and exhibitions, much of the effort lies in coordination.

A client changes the brief. A venue revises its proposal. A speaker needs a different arrival time. Several people need to know, and each needs a different part of the information.

AI can assist with turning meeting notes into proposed actions, comparing supplier quotations against common requirements, preparing handover summaries, and drafting communications in different languages.

Imagine a production meeting that changes rehearsal times and presentation deadlines. Using the approved meeting record, AI could prepare a proposed task list showing the action, owner, due date, and source of each change.

Missing information should remain visibly missing. If nobody agreed who would contact the speakers, the system should flag an unassigned action rather than invent an owner.

A shared workspace could support this approach, provided the team establishes which records are current and approved. Connecting AI to inconsistent notes would leave the underlying coordination problem unresolved.

A useful first application: convert one recurring project meeting into a reviewed action list and team update.

What to measure: preparation time, missed actions, incorrect assignments, and the amount of rewriting required.

Decisions about supplier commitments, expenditure, and operational changes still need an accountable person. A draft should not quietly become an instruction to a contractor.

3. The attendee: getting more value from limited time

Participants face a different challenge: deciding where to spend their attention.

Before an event, AI can help compare published programmes against a stated goal. During it, it can assist with organising permitted notes and identifying questions to revisit. Afterwards, it can help turn those notes into a reading list or follow-up plan.

Take a delegate interested in one particular research method. Given the current programme and the delegate’s preferences, an AI assistant could suggest relevant sessions, explain the choices, and highlight timetable clashes.

The explanation is important. “Recommended for you” gives the delegate little basis for judging a suggestion. “This workshop addresses the method you named and includes practical exercises” is something they can evaluate.

Recommendations also need practical checks. Is the session included in registration? Does it require advance booking? Is there enough time to reach the room? Has the timetable changed?

A useful first application: build a draft personal agenda from the official programme, with a reason for each recommendation.

What to measure: factual accuracy, planning time saved, and whether the chosen sessions actually served the attendee’s goal.

A personal agenda should leave room for discovery. Delegates may want to step outside their usual interests or continue an unexpected conversation.

The same change, seen from three sides

A session moves to another room.

The organiser needs confidence that the change has been approved. The delivery team needs to update signage, staff instructions, and participant communications. The attendee needs the correct location before setting off.

AI might help identify the affected messages and prepare updates. Its usefulness depends on access to the approved change and a clear process for distributing it.

Without that connection, even a fluent answer can send someone to the wrong room.

This is why useful AI adoption begins with the quality of the working information: clear ownership, current records, and a way to resolve uncertainty.

Start with one job you can evaluate

Choose a repeated task with a clear input and an output someone can check. Record how the task is handled today, then compare the AI-assisted version over several attempts.

Ask five questions:

• Whose problem are we solving?
• What approved information will the tool use?
• Who checks the result and handles exceptions?
• What information should not be shared with the tool?
• What improvement would justify keeping it?

Count review and correction time as part of the work. A draft produced in seconds may still create more effort than it saves.

For recordings, delegate information, and confidential documents, establish permissions and appropriate handling before use. UNESCO’s guidance on generative AI, though written for education and research, reinforces the importance of human agency, privacy, and evaluating whether a tool is suitable for its intended purpose. (UNESCO: Guidance for Generative AI in Education and Research)

What comes next

In the next articles, we will explore these three perspectives in more detail: planning and participant support for organisers; shared knowledge and coordination for PCOs and MICE teams; and event discovery, personal agendas, and learning for attendees.

For now, choose one task that repeatedly consumes time at your event. Define what a good result looks like, give AI an appropriate role, and check whether the work improves.

That is where an event starts to become AI-able.