AI champions: building tools and sharing knowledge together

Published AI Braindump

A lot of companies are working out how to use artificial intelligence (AI) better and further. Some are well along, others are still figuring out how to start.

What I hear a lot is that a few colleagues are already working with AI and have built things themselves. But nobody really understands what those things do, and above all: are they safe to use?

What is an AI champion?

An AI champion is someone in a company who works with AI a lot. Colleagues already listen to their tips now and then, and look at the solutions they've come up with. They don't have to be technical, although it helps.

A small team of AI champions can come up with solutions together that move the company forward. Someone more technical builds the tools and checks that it's all actually safe. Someone else looks every day at how things could work better. And when needed, someone keeps an eye on budgets and time, more of a project manager.

Time to try things

Give your AI champions the time to keep going. They know their work, can see what's cumbersome and want to try doing it differently themselves.

Giving people that room to try things matters to me in my own role as an AI champion at VML and KdG. The AI Daily Brief recently covered that role, and I agreed with a lot of what they said. Especially that companies need to be willing to invest in the people who want to get involved.

At VML, alongside the client work, I get room to experiment with AI and new AI models. WPP, the group VML belongs to, has also built its own international platform, WPP Open, which takes some of the operational work off people's hands.

At KdG I work on AI tools for teachers and students. Together with Sam from GeNx, who also teaches there, I build tools to make our classes better and more consistent and to improve the organisation. One example is a platform for our slides and the extra course material around them.

A champion also needs to want to keep learning, because using AI is how you find out what you can do with it. Explaining things to colleagues and working out together why something doesn't work are part of it too.

Atlassian and Dropbox, for example, deliberately set aside time for it. That's necessary, because when it comes on top of the regular work, it's the first thing to slip.

Maintaining tools

A tool that nobody else knows how to run is still a risk, though. If it is used every day and its creator becomes unavailable, who keeps it working?

The team of AI champions can take that on: checking useful tools for safety, maintaining them and developing them further. They don't have to be the people who built the first version. Do agree on who does what, and give those people time for it.

Working with IT

Don't forget to involve IT, because tools like these really shouldn't end up under the radar. IT needs to be open to that too, of course.

Together you can look for the right AI tools and models, with the right guardrails (the limits and boundaries within which AI may be used) and enough security to make sure no data leaks out. IT already has plenty of work, so the day-to-day maintenance doesn't have to sit with them. Microsoft also distinguishes the champion's role from the work of the teams supporting them.

Passing on knowledge

When colleagues improve tools and skills together, they also capture how they do their work. A company develops its own way of working over the years, and that knowledge is often spread across different people.

What one colleague figures out, the next one can use straight away. When someone leaves, you keep more than a tool nobody dares touch. The team can carry on with the knowledge and work that went into it.

New colleagues can build on that too. They learn the team's approach while using those tools and skills, rather than having to discover everything themselves.

Take a look at AI training for businesses if your team wants to learn to work with AI.

That small team looking after the tools could also be me. I build AI agents and automation for businesses and keep them running afterwards, together with your IT team.

The US administration now wants to call AI ‘Super Intelligence’ (SI). Apparently AI has already been promoted. I'm sticking with artificial intelligence for now, even if it takes a little more effort to spell.