What event planners are telling us about the AI bubble

Is AI a sound investment in a game-changing technology or an over-leveraged bubble that's about to go pop?

There’s a chance it could be both.

My eyes were opened to the potentially transformative power of artificial intelligence at the 2023 edition of IMEX Frankfurt. It was the press launch of a generative AI platform that was promising to make life immeasurably easier for event planners by doing their jobs for them.

Afterwards, I was given a demonstration of how quickly it could build out an event programme, based on a random subject of my choosing. Being in a slightly contrarian mood – and wanting to test its mettle - I chose the Life and Times of the Romantic Poets.

Over the next 20-30 seconds I watched agog as the AI churned out a three-day conference programme, with suggested session titles, speakers, plenaries, breakout sessions, and networking breaks. It was a truly disorientating experience. Impossible.

Of course, it would still need a human to sense check and cross reference the entire thing, but, hey, early days. The speakers looked promising (apart from the two or three it had hallucinated) and the schedule looked okay (except for the sessions that were too long) and, oh, it forgot to include a coffee break, and … well I did say potentially transformative …

Since then, corporate investment in AI has surpassed $1.8tn, most of it spent on the infrastructure – data centres, microchips, servers etc., – needed to support it.

Those in the bubble camp say that - despite the jaw-dropping sums involved - AI hasn’t made the kind of gains that justify that level of investment. LLMs are still prone to error and rely on a decent amount of human oversight. What’s more, they say, AI isn't being used to transform businesses. It's being used for basic search and to produce low-grade, ephemeral content. AI slop.

Others disagree, arguing that AI is improving exponentially, that it is far less prone to error than it was, and - as with any truly transformative technology - it is normal for revenues to lag investment. Even if AI proves to be a bubble, they say, it will be a ‘productive’ bubble, meaning - just as the dot.com crash of 2002 didn’t spell the end of the internet - it wouldn’t mean the end of AI.

It's already difficult to imagine a world without AI, that much is true.

Because the likes of Microsoft, Amazon, and Google are amongst its biggest investors, the technology has been foisted on us whether we like it or not. Open a Word document and AI is winking at you. Go into PowerPoint and AI wants to help. Open your emails and AI wants to respond to your messages and manage your workflow. And if event planners are looking for an unusual venue for 350 people? Google’s AI Overview has already done the grunt work.

And yet ahead of the two biggest players in AI going public, a little twitchiness is creeping in. Open AI has pushed its IPO back to next year citing ‘safety fears’, while last week Anthropic investors, perhaps spooked by the company’s slowing growth rate and the costs of AI infrastructure, reportedly shaved $500bn off the company’s valuation.

The bubble question will be settled when it becomes clear whether there is a viable business model underpinning AI. Are businesses willing to pay premium dollar to use ChatGPT, Claude et al for the companies behind them to turn a lasting profit?

The gamble here is on AI becoming the backbone of the economy. For that to happen enough businesses need to have faith in its benefits at an organisational level to pay for subscriptions at a sustainable price. Which in turn will rely on the computing cost of AI getting much cheaper.

Every AI prompt costs X number of tokens, with the number of tokens spent depending on the length and complexity of both the prompt and the response. But most subscribers to Claude or ChatGPT are paying a flat monthly fee, not a fee per token, which means users of Anthropic or OpenAI could be burning through thousands of pounds worth of tokens for a monthly fee of $200. In any event, the vast majority of people using LLMs are still doing so for free.

Does the events industry hold any clues as to how all this will play out?

Event planners represent a discrete group of professional people who have been encouraged to incorporate AI into their working lives. In that sense they are typical of other people working in other industries.

But are event organisations actually committing money to AI, or is it mainly individuals using the free or low-cost versions of LLMs as handy, time-saving tools?

The Soundings AI Impact study, which ran for a year, found AI usage amongst event planners rising from 59 per cent in February 25 to 77 per cent in March this year. That’s impressive growth for an industry not known for being technology adopters, but that figure doesn’t say anything about how the AI is being used or if it has been embedded in organisation workflows.

Scratch the surface a little deeper, however, and the data suggests the biggest take-up of AI is from individuals looking to save time and reduce friction, rather than organisations looking to enhance operational efficiency.

Content creation and marketing were the main areas in which it is being used, the report found, at 65 per cent and 42 per cent respectively. When it came to logistics and operations it dropped to 30 per cent, while only 20 per cent of respondents were using AI for scheduling and resource allocation, a mere five per cent for budget tracking and cost optimisation.

Tellingly, only 42 per cent of participants said AI had made their team more efficient. So even among a population where more than three-quarters use AI, fewer than half reported a clear benefit.

When it comes to investing in the future – the industry doesn’t seem entirely convinced either. Eight out of 10 organisations said they won’t be spending more than 10 per cent of their training budget on AI, while 70 per cent said they planned to spend less than 10 per cent of their tech budget on AI.

Asked about workforce plans, 47 per cent expected no significant change, a third were 'folding AI into existing teams', and only seven per cent were hiring for roles AI has created. Soundings’ own conclusion is that most organisations treat AI as a productivity add-on, not a core capability. The survey doesn’t say which tools people use or who pays for them. A personal subscription and a company licence look identical in a tick box, and plenty of the 77 per cent may be using a free tier. What the data does show is that where budgets are concerned, spending on AI is negligible at best.

Elsewhere the June 2026 Northstar and Cvent PULSE survey of 429 North American planners painted a more nuanced picture, pointing to both individual and organisational use. Sure, 'content creation' came top at 34 per cent, but analysing attendee feedback (33%) and sourcing (31%) were not far behind. RFP evaluation (22%) and predictive analytics (20%) also figured, which suggests some embedding in wider organisational workflows.  

And AI was having an effect on productivity, even if the results were modest. Half of respondents cited time savings, 30 per cent faster execution, and only 17 per cent reported no measurable impact. However just 14 per cent cite pipeline or revenue gains.

Overall, then, the picture is of steady, pragmatic adoption - which neither confirms nor rules out a bubble. However, productivity gains without corresponding revenue gains could support the skeptics’ view that businesses might think twice about paying for organisation-wide subscriptions.

That said, people are unlikely to abandon a tool that saves them an hour or two a week, and nothing here suggests retreat. If a correction comes, it will probably be around expectations, not usage per se. 

The data suggests AI is not fundamentally changing how organisations operate, at least for now. Usage is broadly focused, spending is modest, and the structural changes people talk about - new roles, redesigned teams and operations run entirely by AI - are mostly not happening in the events industry right now.

That might all change if companies integrating AI more fully start to demonstrate a clear competitive edge. If it doesn't - and the events industry is a reliable bellwether for the wider economy - the next thing you type into Claude or ChatGPT might be when, not if, the AI bubble is going to bust.