AI-powered software is changing how businesses work. But to what extent does your business really need it?

When we talk about AI with clients, the conversation often jumps straight to: “Can AI replace this person, or these people, or me?”

Two or three years into the AI revolution, it’s still a question we hear at the beginning of almost every innovation project, from project leaders and the people themselves.

But there’s a complexity scale that we use with our clients to help answer a more important question: "What role should AI actually play in your organisation?" Asking the right questions at the outset helps us find the right position on the scale. Broadly speaking, these 3 zones.

1. Replace your people
Automate tasks entirely where it makes sense.

2. Augment your people
Give them tools that make their existing work easier, faster and better.

3. Supercharge your people
Give them capabilities they simply didn’t have before. Let them tackle problems that were previously beyond the capability of your organisation, or even your industry.

Most businesses we help implement AI with today are still firmly in realms 1 and 2, replacing and augmenting people, increasing their capabilities and improving time and cost efficiencies.

But 3? That's where it gets really interesting.

We're only beginning to see what happens when AI isn't simply used to improve how we do things today, but to make entirely new things possible. And as a designer of software products, this changes the question I have to ask.

For years, the starting point of good UX was relatively straightforward: Who is the user? What problems are they having? What do they need? How can we design software that helps them achieve it?

But agentic AI introduces another actor into the system. We're increasingly designing software that doesn't just help a person do something. We're designing agents that do things on a person's behalf. That means we need to understand not only the human user, but the agent acting on their behalf.

What does the agent need to know?
What decisions can it make?
What should it be allowed to do? Not allowed to do?
When should it ask the human?
How does the human understand, trust and override its actions?

We're creating shadow personas, digital actors representing the goals, intentions and behaviours of our users. And that's a pretty fundamental shift in how we think about UX and design thinking in the AI age.

The mistake is thinking every AI project needs to aim for replacement. Sometimes the biggest opportunity is to make your people capable of more. And sometimes, the opportunity is to do something nobody has done before.

For designers, that means the job isn't simply to design better interfaces anymore. It's to understand how humans, software and increasingly autonomous agents work together. And that's the kind of problem we solve at New Icon, helping organisations understand where AI can genuinely create value for people and the wider business, and design the products and systems to make it happen.