Tapping these internal “AI champions” has become a critical method for driving workforce adoption, and the stakes are high. Companies are spending heavily to remake their businesses around the technology only to find that workers who either don’t trust the tools or don’t know how to use them are among the biggest roadblocks.

“Humans don’t like to change,” said Howard Glazer, co-head of Ropes & Gray’s global private-equity transactions practice and himself an internal “champion” at the law firm. “Saying ‘Here use this new tool’ is scary to people.”

Hesitation can be especially rife in fields like law, Glazer said. “You see these articles about firms being sanctioned because they have hallucinations in their court filings. And the reaction of someone who’s already hesitant is, ‘Gee, I can’t do that,’” Glazer said.

That said, companies are making progress. According to a June report from Boston Consulting Group, 74% of front-line employees, defined as individual white-collar workers with no managerial responsibilities, are now regular AI users, tapping it daily or several times a week. That’s up from 51% in BCG’s 2025 report. The BCG report surveyed some 11,749 employees across 14 markets.

On-the-ground “champions” are playing a key role in those increases, companies say. Through these programs, workers volunteer to receive early access to new tools, special training and opportunities to present to senior executives. In exchange, they’re asked to promote AI adoption to their colleagues and field questions, through both formal meetings and informal conversations.

At Ropes & Gray, they have helped drive adoption of legal AI products like Harvey. Two years ago, 32 Ropes & Gray users sent a few hundred prompts a month to Harvey. Today, nearly 2,200 Ropes & Gray employees generate over 282,000 prompts monthly, and each user is three times more active than a year ago, the company said. The firm said it has about 3,000 total employees globally, 1,500 of whom are attorneys.

A lot of the day-to-day work for a champion like Glazer involves knocking on colleagues’ doors and persuading them to actually try the tools in the first place.

“I go to people that don’t use it, and I have a frank conversation,” Glazer said. He tells them: “It’s going to make the work you do faster, easier, better. It’s just going to be hard at first.”

Recently, Glazer took on an expanded role as “Head of Practice AI,” advising the rest of the firm’s 60-some champions on how to sell their colleagues more effectively. The firm hasn’t formally tracked the success of the champions’ effort so far, but part of Glazer’s expanded role will be doing so.

When colleagues raise concerns over its accuracy and hallucinations, Glazer said the point that always hits home is reminding them that while AI isn’t perfect, neither are the junior professionals who work for them. In essence, they never expect the first draft that hits their desk to be in its final, perfect form.

But one of his biggest tips is that face-to-face, one-on-one conversations are much more effective than group training sessions. In groups, he said, “everyone would just smile and nod and say, ‘Oh yes, we’re using it.’”

While personal conversations spark interest, it’s vital to educate people on using the tool for real tasks they’re working on, said Josh Goldsmith, head of Digital Solutions and Innovation for Internal Audit and an AI champion at Citigroup.

For example, Goldsmith worked on an AI use case that used the firm’s internal AI platform, Citi Stylus Workspaces, to generate succinct drafts of audit reports. Selling colleagues on that specific capability “has gone a long way in changing people’s minds,” he said.

The bank has sizable AI ambitions for its workforce. At Citi’s investor day in May, Chair and Chief Executive Jane Fraser said it is applying AI across the business, from accelerating client onboarding and underwriting processing to building AI-powered virtual wealth advisers to serve clients, as well as using it in cybersecurity and coding.

To help meet those goals, Citi has scaled its program to more than 4,000 participants, divided between “champions”—or senior-level executives—and “accelerators,” those at various levels of the organization.

Adoption of Citi Stylus Workspaces, the most broadly available AI tool for workers at the company, has crept up from the single digits in late 2024 when it was rolled out to more than 80% now. The company said it credits the champion and accelerator program in part with that increase.

Another key to the success of champion programs is making sure you tap the right people. According to Mars Snacking Chief Technology Officer Ramesh Kollepara, the most effective AI champions “are definitely not the most technical people.”

Rather, you’re looking for someone with innate curiosity who can act as a translator between business needs and technical capabilities, he said.

Kollepara started a champions program two years ago while he was leading tech at Kellanova, and it has expanded since the snack company’s acquisition by Mars last year, now boasting 500 participants.

“Champions are the ones who socialize the broader potential of AI,” he said.

Write to Isabelle Bousquette at [email protected]