A two-minute read from We Are Zenith on turning early adopters into AI champions

Every organisation has early adopters: people who could become your AI champions.

They spotted what AI could do months ago. They’ve quietly built a few shortcuts: a prompt that drafts the weekly report, a quick way to summarise long email chains, a spreadsheet that half-automates the rota. They do most of it in their own time, with whatever tools they can find, because nobody asked them to and nobody said they couldn’t.

Sociologist Everett Rogers found that innovators and early adopters make up roughly 1 in 6 people. They try new technology first, and they’re your best chance of making AI work across the whole organisation.

Left alone, they’re also your biggest unmanaged risk. Microsoft and LinkedIn found that 78% of people using AI at work bring their own tools, and only 39% had been given any AI training by their employer.

The good news? Turning lone experimenters into AI champions doesn’t take a big budget. It takes three things: a mandate, a method and safe tools.

1. Give them a mandate

Right now, most early adopters treat AI as a side project. Nobody has asked for it, so nobody makes time for it, and nobody else in the team knows it exists.

A mandate changes that. It says: this matters, it’s part of your job, and we’re behind you.

In practice, that means:

• Name it. Tell your early adopters, and their teams, that they’re leading AI improvement in their area.
• Protect the time. Agree a regular slot in the working week, so it stops happening on lunch breaks and evenings.
• Give a clear brief. For example: “Find two processes in your team worth improving, prove the time saved, and share what works.”
• Choose for influence, not just enthusiasm. The best AI champions are people colleagues already respect. The rest of the team will copy them.

2. Give them a method

Most early adopters start with a tool and ask, “What can this do?” It’s exciting, but it rarely changes how the wider team works.

AI champions start with the work instead. A simple method looks like this:

1. List the everyday tasks. Aim for at least ten: the rekeying, chasing, checking and searching that everyone grumbles about.
2. Score them. Rate each one for how easy it would be to improve and how much it would matter.
3. Time the top three first. Measure how long they take today, before anything changes. Without a “before”, any saving can only be claimed, never proved.
4. Estimate, then measure. Write down the expected saving, put the improvement live, then record what it actually saves.
5. Keep a record. Log every idea from first thought to signed-off result, so you build a body of evidence rather than a list of anecdotes.

That last point is the one that matters most. The majority of your people won’t change how they work because an enthusiast says so. They’ll change when a colleague shows them something that works, with the numbers to back it up.

3. Give them safe tools

On a busy afternoon, trying to save an hour, an early adopter pastes a client document into a free chatbot. Nothing goes wrong. But nobody would know if it had.

That isn’t a discipline problem. It’s what happens when keen people have no approved way to do the job properly. Banning AI won’t fix it; it just pushes it further out of sight.

Instead, give your AI champions:

• Approved tools. Agree which AI tools people can use for work, and make them easy to get.
• Clear data rules. Spell out what can and can’t go into an AI tool, in plain English.
• A check before go-live. Any improvement that touches real work gets reviewed before it’s switched on.
• A line on decisions. Be clear about what stays with people. AI can draft the proposal, but a person sets the price.

Do this and shadow AI becomes something better: a demand signal you can channel, with your champions leading the way.

From AI champions to the whole organisation

Most AI rollouts stall in what’s called the chasm: the gap between the keen early users and everyone else. Your early adopters are the bridge across it, but only if you back them.

A mandate gives them permission. A method gives them proof. Safe tools give them confidence, and give you peace of mind.

So here’s a question worth asking this week: who are the early adopters in your team, and what would they need to lead the change?

Want a structured way to build AI champions?

Our Level 4 AI & Automation Practitioner apprenticeship is built for your early adopters. Over 14 months, with one 2 to 3 hour workshop a month, practitioners work through our five-step SCALE model on real processes in your organisation. Each one takes at least two improvements live, governed and measured, and signed off by their manager.

It’s funded through the apprenticeship levy or government co-investment.

Sources

• Everett Rogers, Diffusion of Innovations (1962): innovators and early adopters make up about 16% of an idealised adoption curve.
• Geoffrey Moore, Crossing the Chasm (1991).
• Microsoft and LinkedIn, 2024 Work Trend Index: 78% of AI users bring their own tools to work; 39% have had AI training from their employer. Global survey.