Only 35% of UK businesses report using at least one AI technology, according to the latest Office for National Statistics figures, up from just 12% three years ago. That sounds like a genuine shift, until you read the fine print: the ONS itself admits the figure treats a one-off chatbot experiment the same as AI properly built into how a business runs, and its own data on intensity of use barely moved, from an average of 1.4 AI tools per adopting business to 1.6. If you work in a small events team, that gap probably won’t surprise you. The surprising thing would be the opposite.
I’ve spent months watching conference panels and articles talk about AI for events as if it were an entire department with its own budget, a team of data scientists and a three-year roadmap. Then I get back to my desk, with my to-do list for the day, and the real question is much smaller: which tool saves me half an hour today that I can spend talking to a client or closing a partnership? That gap between the discourse and the day-to-day is exactly what this article is about.
And it’s not an isolated case. In the events sector, agencies with large departments sit alongside corporate event managers who are one or two people handling logistics, registration and communications for the entire event. Writing only for the first group leaves out most of the people who actually organise professional events every week. This article is for that majority.
What gets sold as AI for events, and what actually gets used
An industry report finds that 95% of event professionals expect AI use to grow within their organisation, with 35% expecting a significant leap. The expectation is clear. But the same report acknowledges real doubts about accuracy and data governance: plenty of expectation, not much infrastructure behind it yet.
That’s the problem with a lot of what gets called AI for events today: pilots shown off at a trade show, screenshots of a pretty dashboard, a case study that’s never repeated because it depended on three people working on it full-time for two months. That’s showcase AI. It’s good for the photo, not for Monday morning, when you have to get a 400-person registration list sorted with nobody but you to check it.
What actually works is a lot less flashy. It’s a tool you already use — your event management platform, your CRM, your email marketing — with a layer of AI for events that solves one concrete, repetitive task. You don’t need to reinvent the process. You need that process to stop eating three hours of your week.
DSIT’s own AI adoption research names limited AI skills and expertise as the biggest barrier for UK businesses — cited by 60% of firms overall, and 68% of small businesses that are planning to adopt AI but haven’t yet. Adoption also varies sharply by region: London leads at 20%, against a 16% national average, while businesses in Scotland are far more likely to say they have no plans to use AI at all. In a one- or two-person events team, that skills gap doesn’t get solved by hiring someone else, because there’s no budget for that — it gets solved by choosing tools that don’t require a technical profile to get value from.
What doing AI for events means when the team is just you
If a tool doesn’t give me back time within a week of implementing it, I don’t keep it, however good it sounds in the vendor’s demo.
That changes quite a lot which AI projects I pick and which I drop. I’m not interested in training my own model or building a predictive analytics layer from scratch. I’m interested in the platform I already use solving, with AI, a task I currently do by hand — without me having to learn to code to get there. And with clients running events with equally small teams, this is exactly the angle that matters to them when they ask about AI for events: they don’t want a lab, they want one less task on their list.
All the AI for events worth having in a small team meets that condition: it saves real time, not just looks good in a demo. With that in mind, these are the cases I see actually working for clients with small teams.
Real cases, not showcase promises
1. Generate the first draft, never the final copy
When you need to launch an event landing page, the confirmation email and three social posts in the same week, AI saves you the blank page: it generates a first version of each piece and rewrites it in your voice, with your data and judgement. What doesn’t change is the review: every fact, every figure, has to be checked before it goes live, because a confidently generated but wrong figure costs more reputation than it saves in time. This [article on eight practical AI for events applications](https://eventscase.com/blog/ai-for-events-8-practical-applications) covers the rest of the range well, from audience segmentation to automatic session summaries.
2. A chatbot that filters before it escalates
Before an attendee sends an email asking what time the event starts or where to collect their badge, a well-configured chatbot answers that on its own, twenty-four hours a day. Eventscase builds this logic into EVA, its WhatsApp assistant, and for a small team the difference isn’t philosophical — it’s the repeated questions that no longer land in your inbox. This article on using AI and WhatsApp for event management with EVA sets out the concrete use cases, from travel logistics to QR-code networking.
3. Reading the feedback nobody has time to read
After an event with several hundred responses to a satisfaction survey, nobody on a one- or two-person team sits down and reads them all one by one. Automated analysis of that feedback (grouping recurring complaints, spotting which session landed best, what logistics went wrong) turns a folder of open-ended answers into three or four actionable conclusions before the closing meeting with the client.
4. Measuring whether the efficiency translates into real ROI
The part most often forgotten when people talk about AI in events is the metric. Automating a process is fine, but it only matters if it turns into something you can defend in front of leadership: lower cost per lead, more opportunities per demo, fewer team hours per event managed. This article on event ROI and using AI for efficiency goes into the detail of connecting the two without losing sight of the business.
5. Personalising the agenda without knowing every attendee individually
At a conference with three hundred people registered, nobody on a small team has time to review each attendee’s profile to recommend which sessions suit them. A system that cross-references registration data with the programme content and suggests a personalised agenda does that work in seconds — and it also tells you a lot about which sessions genuinely interest people before the event starts, not just afterwards. It’s the kind of insight that used to belong only to events with big budgets, and is now within reach of anyone running a mid-sized one on tight resources.
The test for deciding which AI project is worth it
Before approving any new tool, three questions, in this order: what concrete task stops being done by hand? How soon do I see the result? And what happens the day the tool fails or gives a wrong answer? That last one gets asked the least and matters the most when you’re working without a safety net: if a chatbot gives a customer the wrong information, the responsibility is still yours, not the AI vendor’s.
When someone searches for AI for events on Google, they’re not looking for an academic paper or a trade-show demo. They’re looking for the same thing everyone is: a short list of things that actually work, who’s tried them, and what went wrong the first time.
There’s something else behind that third question, and it has more to do with trust than with technology. I’d rather work with a vendor who admits where their tool falls short than one who promises it never gets anything wrong. With AI, as with any partner, speed and price matter, but what decides whether I keep working with a tool long-term is whether I can trust what it tells me when something goes wrong.
Where this is heading
The gap between the AI that businesses say they’re using and the AI that’s actually built into how they work won’t close with more panels or more trend reports. It’ll close, event by event, with small teams trying one specific tool, measuring it for a month, and deciding whether it stays or goes. That’s the kind of AI for events worth writing about: the kind you can repeat next Monday, not just the kind that looks good on a slide.