Every conversation about AI eventually comes back to cost.

How much does a Copilot licence cost? What will our agents cost to run? How much consumption are we going to burn through? And how do we stop people spending money on AI without creating any value?

They’re reasonable questions. But they focus on the cost of running AI rather than what AI is doing to the economics of work.

AI is making content, analysis and recommendations dramatically cheaper and faster to produce. What used to take hours can take minutes. What used to take minutes can happen automatically. But making work cheaper to create doesn’t necessarily make an organisation more productive. When writing a report becomes almost free, we can produce more reports. When presentations take minutes, somebody still has to review them. When everyone can generate professional-looking analysis and recommendations, somebody still has to decide which ones deserve attention.

In this session, we’ll explore the economics of AI at work through three lenses: licence economics, consumption economics and attention economics. We’ll look at when paying more for AI can create extraordinary value, when cheap AI becomes surprisingly expensive, and how to decide what work belongs with people, copilots, agents and increasingly autonomous AI.

We’ll also look at one of the less comfortable consequences of AI adoption: sometimes we haven’t removed the cost of work at all. We’ve moved it. A report that takes two minutes to generate can still take ten people twenty minutes each to read. Poor or unnecessary AI output doesn’t just waste the creator’s time. It consumes the attention of everyone who has to read it, review it, validate it, make a decision about it or clean it up afterwards.

Drawing on lessons from real-world Copilot deployments, adoption programs and customer conversations, this session gives business leaders a practical way to think about AI investment beyond licence counts, usage statistics and token consumption. The goal isn’t to spend as little as possible on AI. It’s to understand where spending more creates value, where AI is simply creating more work, and what’s actually worth doing in the first place.

Because the most expensive prompt in your organisation isn’t necessarily the one that costs the most to run.

It may be the one that creates work for everybody else.