Event planners should stop trying to keep up with AI

 

Nobody ever tells an event manager they’re falling behind because they haven't learned to code. We don't expect everyone who uses the internet to know how to build websites or create applications. Some people develop extraordinary capabilities with internet technology because doing so is useful to them. Most of us just use the bits we need. 

So why do we treat Artificial Intelligence differently?

Why has the events industry developed this strange expectation that everyone should continuously advance towards ever-more sophisticated AI use?

Not long ago, I ran a (very) unscientific poll in a WhatsApp group of 642 freelance event professionals. I just wanted a sense of how people were actually using AI.

Seventy-three people responded: three were using agents and MCPs, seven weren't using AI at all, and the other 63 (86%) were using it for emails and relatively basic tasks.

It is a tiny, self-selecting sample and I wouldn't pretend it tells us anything definitive about AI adoption across the industry, but I found it interesting, nevertheless.

If you were only going by the noise around AI, would you expect those numbers?

There’s an awful lot of noise around AI in the industry at a moment - and a persistent message that if you aren't getting to grips with all of this, you're going to be left behind. But the noise around a technology isn't necessarily evidence of the extent - or sophistication - of its use.

Again, the internet is one of the most extraordinary technologies ever built, but most of us use it to send emails, read the news, book holidays, buy things and look up information. It can empower us to do more than that, but most of us don’t want or need it to do more than that.

There is no inherent virtue in moving up an imaginary ladder of AI sophistication just because. If you have a complex problem worth solving, improving one’s capability in AI can make sense. But if what you need from AI is help writing emails, or marketing copy, perhaps that's enough.

There are extraordinary things you can do with AI, of course, and the speed at which they appear to happen is phenomenal. But that capability to respond to a prompt in a split-second is the result of huge computational power and decades of human input, shaping, and refining. It might feel effortless, but, upstream, an awful lot of effort has gone into it. 

That distinction gets lost in much of the conversation about AI.  So, we end up talking about what these systems are capable of as though the capability itself creates value. But meaningful use requires something from us first: an understanding of the problem, and knowing enough to judge whether the output from AI actually has any merit.

But what about increased productivity?

Much of AI's allure comes from the idea that there is something inherently virtuous about increased productivity. There isn't.

Recent research that looked at 100,000 developer activities found that autonomous coding agents increased coding activity by around 180 per cent, but software releases only increased by 30 per cent. Similarly, across four major app marketplaces, there was a 180 per cent increase in new apps, yet no increase in total downloads. We may have increased app-production output, but we have not developed a corresponding appetite for more apps.

That’s because productive capacity and value aren't the same thing. Just because we can produce ten times as much does not mean that the world needs ten times as much or is even capable of consuming ten times as much. And if there isn’t a corresponding appetite for increased outputs, then there is no value associated with those additional outputs.

An event team can now produce more marketing variants, more personalised emails, more session summaries, more attendee recommendations, more reports and post-event content than it could before. But a delegate hasn't acquired any more hours in their day.

They can still attend only so many sessions, read so many emails and absorb so much information. There is a limit to the number of meaningful conversations one person can have.

This truth creates a tension between the economics of AI, which makes abundance incredibly cheap, and the economics of human attention, which remains profoundly scarce.

So before celebrating the fact that AI has enabled us to produce 100 things where previously we could produce ten, we should ask ourselves: did anybody need the other 90?

Asking that question is critical in decided whether to use AI to achieve our needs. It also helps us decide if using AI warrants the cost - both economically and environmentally.

AI is a tool. Let’s not forget that. But no-one uses a hammer when they want a paintbrush. 

The default position should not be to treat AI as the default tool, but to start with the problem at hand: “Here is something we want to make better; what could help?”

Because the answer might be a person, it might be a new process, or it might be AI. 

Deciding on when and how to use AI is important, because AI doesn't contain its own stopping condition. If it makes something twice as efficient, nothing within the technology tells it to stop, or encourages us to take the efficiency saving and stop, even when producing more stuff has no value. This brings into consideration the environmental cost of AI.

Behind every response sits an enormous physical system: data centres, semiconductor manufacturing, electricity generation and transmission, cooling infrastructure, water, minerals, land and networks. Ellie Ashton-Melia has written brilliantly about the environmental costs, but I'm interested in the question immediately before that one.

What is all this computation actually for?

I don’t think every prompt needs to justify its existence against a carbon ledger, nor do I think there is an argument that effort should be preserved simply because effort is virtuous. There is plenty of work that is tedious, repetitive and necessary (filing tax returns, reconciling budgets, cross referencing attendee lists and room allocations), and if machines can do it while people do something more worthwhile, that can be an excellent use of technology.

But the issue is that AI removes effort, one of the key constraints that historically limited how much we produced.

Time, expertise, labour and money have always acted as filters on production. They don’t guarantee that everything we make is useful, but until now producing something has required enough effort or investment for somebody to decide it is worth doing.

AI dramatically lowers that threshold. And because the physical cost of the technology is largely invisible to us as users, we can lower the perceived cost at both ends simultaneously. It requires less human effort (cost) to produce something, while the resources (environmental capital) consumed in enabling that production happen somewhere else, out of sight.

That combination makes abundance extraordinarily easy, but it doesn't give us any better mechanism for deciding what deserves to be abundant.

And this is where I think the environmental question becomes inseparable from the productivity question. If additional economic output requires additional computing infrastructure, energy and resources, then "we can produce more" is not sufficient justification for producing more - we need some conception of the value being created.

If AI helps an event team remove hours of repetitive reconciliation work, improves accessibility for delegates, identifies something important in data that humans would have missed, or gives a small organisation a capability it previously couldn't afford, there is something tangible to put on the other side of that equation. If it allows us to generate another 500 pieces of content that almost nobody reads, the case is rather harder to make.

Which brings me back to those 73 event professionals. It would be very easy to look at the poll and conclude that our industry has an AI skills gap - and perhaps we do. There will certainly be event professionals that could get more value from these tools than they currently do. 

But that isn't the only interpretation.

Perhaps some of those 63 people using AI for emails and basic tasks have simply found the level at which the technology is useful to them. Perhaps using enough AI is enough.  

The three people using agents and MCPs aren't necessarily "ahead" - they are using different capabilities. Whether those capabilities represent greater progress depends entirely on what they're achieving with them. Selective use of technology and AI doesn't mean resisting it, it means preserving the order in which we ask the questions: what are we trying to improve? Why are we trying to improve it? What is currently stopping us doing that? Could a technology help?

Sometimes the answers will justify sophisticated AI systems and significant investment. Sometimes they will justify asking Claude to tidy up an email. Sometimes doing the thing yourself will remain more useful, more enjoyable or more important because the process is where you develop the judgement you will later need. And sometimes the most useful conclusion might be that the thing didn't need doing in the first place.

So, the next time you feel like you’re getting left behind remember, there is no professional obligation to automate everything that can be automated, or to generate everything that can now be generated. The useful measure of AI maturity isn't how much of it an organisation uses, or how sophisticated that use appears from the outside, but whether the organisation can explain what has become better by using it.