Event technology trends in 2019

How to read this article

This article was written in June 2019 and is kept as a record of what organisers were discussing then. It is not a current trend report. For the present-day view of chatbots and attendee communication, see the pages linked at the end.

The landscape moved. The meeting still has a pipeline.

AI, social platforms and virtual rooms changed the surface of events. Organisers can message attendees in Messenger, stream a keynote, or answer FAQ without printing a brochure. None of that replaces abstract intake, peer review or a timetable chairs can defend. Eventumbot is conference management software first; the event chatbot is how communication keeps up with that record.

AI is useful when it reads real data

A bot that cannot see the live agenda will hallucinate room numbers. Machine learning helps when profile, location and FAQ sit on the same event as the scientific programme. That is the Eventumbot bet: automate the predictable questions, escalate the rest to the info desk.

Social is a channel, not a source of truth

Attendees already live in messaging apps. Using those apps for notifications is sensible. Using them as the only programme is not. The public event website and the conference app should publish the same sessions the chatbot quotes.

Virtual and hybrid still need operations

Virtual rooms did not delete fee tables, authorship or conflicts of interest. If anything they made last-minute changes more visible. Conference software that already joins intake, review and the timetable is in a better position to add a chatbot or a stream link than a communications widget that never met the committee.

What organisers were asking in 2019

When this article was written, organisers were asking whether chatbots, social platforms and virtual rooms would change their work, and how much of their budget these things deserved. The questions were practical: can a bot answer attendees so the info desk is free for harder cases? Can we stream the keynote? Should we build an app? The excitement was real, and so was the risk of buying a technology because it was new, not because it solved a problem the meeting had.

Which ideas have held up

Looking back, several ideas from that period remain useful because they rest on organisers’ real needs, not on any single product.

• Answer predictable questions automatically, and hand over the rest to people.
• Meet attendees in the channels they already use instead of asking them to learn a new one.
• Keep one source of truth for the programme, so that every channel shows the same thing.
• Treat virtual and hybrid as additional formats that still need the same operations: fees, authorship, conflicts and a timetable.
• Measure what happens after the meeting, not only during it.

What has changed since

The tools have moved on. Software that answers questions in natural language has become far more capable, and attendees’ expectations of a fast, accurate answer have risen with it. Messaging platforms change their features and rules over time, so a channel that worked well in one year may work differently in another. Virtual and hybrid formats are now familiar, and organisers are more critical about when they add value. What has not changed is that a tool is only as good as the information behind it and the humans who look after it.

How to judge a technology claim

Vendors and articles, including this one, describe technology in its best light. A few questions cut through it.

• What problem does it solve for our meeting, and how will we know?
• What information does it need, and who keeps that information up to date?
• What happens when it does not know the answer?
• What does it cost in staff time, not only in price?
• What will we lose if we do not use it?

Questions to ask a supplier

Whether it is a chatbot, an app or a streaming service, ask the supplier for concrete answers before you commit.

• Can I see a real event that used this, and speak to its organiser?
• Which parts are configuration and which are custom work?
• Where are the data stored, and who can see them?
• How do I export my data at the end?
• What support is available on the day of the meeting?
• What are the limits of the service, in plain words?

A framework for small experiments

If you want to try something new, run a small experiment at a meeting that can afford it. State the problem, define what success would look like in numbers, set a limited scope and decide in advance what you will do with each outcome.

Problem: the info desk answers the same 20 questions all day. Experiment: add an FAQ chatbot for the two-day symposium. Success: at least half of the questions answered without staff; no wrong answers about room or time. Check: read the conversation log each evening. Decision after the meeting: keep, improve or drop.

What this means for the conference pipeline

The technology on the surface of an event changes faster than the work underneath it. Submissions, review, the programme and the registration still need an owner, and the communication tools depend on them. That is why we treat the scientific record as the foundation and the chatbot as a way to communicate it. A new tool is worth adopting when it makes that record easier to understand for attendees, and it is worth avoiding when it creates a second version that nobody maintains.

Using this retrospective in a planning meeting

Old questions are still useful prompts. Before your next committee meeting on technology, ask each member to write down one channel or tool they think the meeting needs, and the problem it would solve. Sort the answers into problems that occur every year, such as repeated attendee questions, and problems that are specific to one edition. Then check each proposed tool against the questions above. Most of the time you will find that two or three of the ideas address the same recurring problem, and that one of them can be tested cheaply. Write the result in the decision log so that the reasoning is available to whoever chairs the next edition, when the tools will look different again. Keep the exercise short: thirty minutes is enough, and the goal is a shortlist, not a purchase decision.