“The problem’s not getting the tools to the people. It’s getting the people to the tools.”

So spoke one of our clients recently, echoing the sentiments of every techy who ever found themselves in the uncomfortable position of being ahead of the curve.

In this case, he’d built a private AI tool that rewrites comms in his company’s tone of voice. But the story’s millennia old.

Email as a business tool? Madness at first (it’s not secure! It’s not private! It’ll distract people! It’ll get in the way of my two-hour lunches!), everyday now.

Mobile phones, computers, calculators, typewriters… Plato even said this about a pesky fad that would surely never catch on:

“If people learn this, it will implant forgetfulness in their souls. They will cease to exercise memory because they rely on that which is written, calling things to remembrance no longer from within themselves, but by means of external marks.”

What was he talking about? He was talking about writing, ladies and gentlemen. WRITING.

It’s safe to say no matter how transformative AI is, your people won’t necessarily just adopt it because it’s right there in front of them.

It’s not a tech project. It’s a culture change project.

Business leaders are terrified about data security. Employees are terrified of losing their jobs. Everyone’s terrified of change, full stop. And if you don’t factor in those fears and foibles, you’re facing a rocky road to adoption.

So what to do?

Here’s the three-step process we used to launch our own private AI (based on decades of experience launching things lots of people don’t want, like tones of voice). And it’s the process we encourage our clients to follow too.

1. Start with the hard questions

At some point, someone’s going to stand up at the annual meeting and ask the CEO if AI is going to take everyone’s jobs.

So involve your people in the process of finding answers to all the uncomfortable questions.

Send out a survey. Run workshops. Do research. This is your chance to build a solid foundation for your AI, and make sure your people feel involved (because the IKEA effect).

2. Dip your toe

Choose a small group of guinea pigs, train them to write brilliant prompts, and give them a sandbox AI to play in. It means you can show off how powerful it is quickly and cheaply.

There’s a sneaky psychological reason to do it to, based on the scarcity principle: giving such a powerful tool to a select few means everyone else will start to want it.

It’s also worth using this pilot to gather some success stories. Have the guinea pigs seen tangible, measurable improvements in their performance? Great. Jot those down. You’re going to need them later.

3. Go all-in

The foundations are laid. The buzz has spread. You’re ready to bring AI to everyone.

With a private AI powered by APIs, you pay for use, not users, so you can go from sandbox to Sahara in seconds without needing to sort out hundreds of expensive subscriptions.

Of course, there will still be naysayers. So follow the COM-B model of behaviour change and focus on giving people three things: the capability to change, opportunities to change, and the motivation to change:

For capability: run plenty of prompt training to help your people get the most out of it.

For opportunities: create built-in prompts for anything from press releases to social posts to your tone of voice.

For motivation: run employee engagement campaigns, driven by the Technology Acceptance Model (in a nutshell: if you can prove your AI is useful and user-friendly, people will use it).

3½. Relax

You just made your entire workforce more efficient, creative and productive. And all it took was a little help from your friends at Definition.


Give us a shout to talk about starting your own private AI journey; we’ve got all the tech types, writers, designers, employee engagement consultants and process experts you need to do it right.

Nick Padmore Screen

Written by Nick Padmore, Head of Language at Definition.