At software maker ServiceNow Inc., learning and development executive Jayney Howson leads a small team responsible for getting 28,000 employees trained on artificial intelligence over the next three years.

Rather than go it alone, her group identified 1,000 high performers from around the company, all of whom were known to be avid AI users, and offered them corporate swag and recognition in exchange for teaching their colleagues the technology. Howson’s team gives them facilitation techniques and tips on how to inspire a room, “and then we kind of leave them to it,” she said.

Other employers in industries including tech, finance, manufacturing and pharmaceuticals have landed on a similar approach, recruiting influencers from their own ranks to help spread AI know-how. The idea is to inspire people to use AI, instead of trying to force them into it with blanket mandates to use the technology, or threats to their job security if they don’t.

Executives who’ve tried the influencer model — though they usually don’t refer to it that way, instead using labels like “ambassadors” or “champions” — say it’s effective at addressing employees’ concerns, mistrust or sheer lack of interest in the new tools, and may help close the AI enthusiasm gap that often exists between executives investing in the technology and the rank and file who are expected to adopt it.

“I can all day say, ‘AI matters, we should be using it, it will drive our impact farther.’ Hearing an executive say it is really different than hearing your peer be like, ‘Oh, actually I can get my work done 30% faster and then I can focus on these other things,'” said Bijal Shah, chief executive officer of workforce education company Guild, which assembled an internal team of AI “activators” from across departments to help spread the word.

Some companies are turning to the influencer strategy after earlier, top-down adoption efforts stalled out. At commercial vehicle maker International Motors LLC (formerly Navistar Inc.), traditional training worked for those keen to experiment, but wasn’t effective for the large majority.

“We started with training like everyone else; it didn’t work,” said Tim Foley, International Motors’s head of AI adoption. “You think this would sell itself, but it doesn’t. People need 10 hours of practice and they’re so busy.”

So Foley devoted time to helping early adopters “find their voice” to become influencers. They’re given opportunities to work on their messaging, bots that help them write social posts and even premium licenses to AI tools, which they can share with colleagues of their choice.

How Influencers Surface

At Morgan Stanley, Alisha Lehr, chief operating officer for firmwide AI, says she scouts influential and intellectually curious vice presidents from across the company to join her “AI ambassadors” team. Those VPs are deputized to run division-specific training. Some have created cascading networks of influencers within their own groups. “Scale is key,” she said.

Other companies designate their in-house experts less formally, perhaps by maintaining internal leaderboards that feature the heaviest users of AI tools, or through subtler means, like inviting employees to showcase their AI workflows to colleagues at all-hands meetings.

Yelp Inc.’s vice president and chief of staff of engineering, Yoann Roman, says it’s pretty easy to identify potential influencers by talking to department heads and keeping an eye on internal Slack channels devoted to sharing AI tips and use cases, where those passionate about the technology tend to congregate.

The enthusiasm and expertise freely shared in company chatrooms can be hard to replicate with the typical top-down approach to employee training, which is where peer-led instruction comes in. Relinquishing control over exactly what’s getting taught in the AI training sessions at ServiceNow made executives nervous at first, Howson said. But letting go was necessary for the scale and speed of what they wanted to accomplish.

Room for Skeptics

Software maker HubSpot Inc. has multiple strategies for deploying in-house influencers, including one inspired directly by social media. Every Monday, the company posts a new “MondAI minute,” a 60-second clip featuring a clever AI use case from an employee who might work anywhere within the company.

“We’ve got a TikTok generation,” said Chief People Officer Helen Russell. “We also have to talk to the organization in a way that they want to receive information.”

Another effort came from the software maker’s customer-support department, which carefully selected 10 employees who held sway with their peers to become AI “champions.” Along with training their colleagues, they meet a couple of times a week to generate readouts for senior leaders on how it’s going.

Key to the strategy: Alongside AI enthusiasts, it recruited some doubters. “We purposely tapped the skeptics because they’re the ones that are like, ‘Here are all of the reasons why it can’t work,'” Russell said.

Making space for that honest feedback built trust, Russell said. It also helped the department surpass its goal for AI adoption. While the AI champions team hoped to help deliver an adoption rate of 70%, they hit 80% in a month, Russell said.