A leadership expert's guide to navigating AI

Gillian Davis is an expert in leadership during periods of change. Through her OverTime Leader consultancy, she’s worked extensively with design, creative and tech teams, helping them navigate difficult and disruptive moments.

While she is convinced AI will have a transformative effect, she also sounds a note of caution about over-confident declarations. “No-one out there can predict the impact and value AI will have, because we’re all just figuring it out.”

Given that many design leaders feel overwhelmed by the pace of change, I wanted to find out what Gillian thinks people should know, and do, during this great uncertainty.

We discussed:

• The two modes leaders should switch between.

• How to actually create a culture of experimentation.

• Why information vacuums are terrible for morale.

Because nobody really knows what’s going to happen, Gillian thinks leaders should adopt two different leadership stances, depending on the situation.

At times, they should be directive, and make a call based on what they do know, and where they want the team to go.

“In a world where nobody knows the answer, it’s going to feel uncomfortable putting a stake in the ground. But there will be times when your team needs you to say, ‘Based on today, the stake in the ground is going to go here, and that’s what we’re going to rally around.’

“You might find out you need to move it two inches to the other side, but at least you’re giving them a starting point.”

Once that stake is in the ground, she says, leaders should become more facilitative. Empower people to explore, invite different inputs, and decide the next way forward together.

Crucially, don’t get stuck in debate loops that never lead to decisions. “Doing is better than talking in this environment.”

To get things moving, Gillian thinks teams must “move into a mode of experimentation.” This sounds great in theory, but can seem quite overwhelming in practice – it’s not very helpful to tell teams to “play around” with AI tools, or encourage them to use up tokens (as that gets very expensive).

Instead she says, get people to write down all the things they don’t like about their jobs. Or all the internal processes they think could be more efficient. Then pick one or two and run an experiment to see if AI can help. Give them a deadline, and follow up to find out how it goes.

Leaders should also be running their own experiments. This sets a good example, but more importantly, it helps leaders understand what’s realistic, what’s possible within user licenses, and what AI tools are actually good at.

“Just because AI can do a thing, doesn’t mean it should do a thing.”

In a culture of experimentation, everyone, leaders included, needs to be very open about what doesn’t work.

“I know it sounds a bit cringe, but role modelling is very important. Good leaders aren’t afraid to share their failures, their use cases, and learnings. That openness gives everyone permission to get into that experimental mode.”

In survey after survey, we see a yawning disconnect between leaders’ enthusiasm for AI, and everyone else’s. The problem often lies in the unrealistic expectations of the former, which creates frustration and disillusionment in the latter.

“Leaders keep making these huge promises and people on the ground are like, yeah, but that’s not possible,” Gillian says. “So my advice to leaders is, before you get up in front of the team and say anything about AI, come back to earth, understand what you’re asking, and what the roadmap to implementing that would actually look like.”

Leaders also need to understand that pursuing any ambitious vision will involve “a bunch of boring work.”

“At so many companies, their data is just a mess. Smart leaders will invest in getting that cleaned and set up properly, and they’ll see the returns a few years from now. But that will cost money, and nobody gets very excited about it.”

Gillian believes that, for the next few years at least, change will be a constant. Leaders need to accept that, and make sure “they’re not trying to build for stability in a world that doesn’t exist anymore,” she says.

They also need to help their teams let go of that idea of stability. Leaders need to be empathetic – some people will find that scary, distracting, maybe even overwhelming.

Organisations need leaders who are themselves comfortable with change, and who can help others become comfortable with change. Gillian thinks you can teach this ability to navigate uncertainty, but it’s quite different “to what’s previously been taught, and rewarded” in most organisations.

In times of flux, leaders will need periods where they’re working on what to do next. What they mustn’t do is disappear into planning mode, so teams don’t know what they’re meant to be doing.

“Communicate constantly,” Gillian says. “Don’t wait for the strategy to be finished – show people where you’re currently at. The longer teams are made to wait, and live in the abstract of what might happen, the worse it will be, with people guessing, over-analysing, and maybe even leaving.”

Whenever someone in Gillian’s team shares a document, they’re asked to make clear how much AI was used in putting it together. The tone of her feedback is shaped by this percentage.

“If you wrote 100% of it, I’ll be more considered in my feedback. If you tell me ChatGPT wrote 90% of it, I’m happier to rip it apart.”

“This is my most passionate point. Do not forget who you’re designing for, and why,” she says. AI creates endless possibilities. Good leaders need to constantly sense check what the team is doing.

“What problem are you actually solving? Go and talk to users. Talk to clients. If you work in-house, go and talk to your sales team, or even your IT team.

“The best use of time in a period of uncertainty is to build non-transactional relationships, based on curiosity. Ask questions. Listen. Don’t come in with a point of view, come in to talk, human to human. The more we talk to people, the better our work will be.”