Journalist and AI workshop leader Jeff Wilser on bringing AI into a company on purpose and keeping people at the center of it.
Ask me whether I’m excited or anxious about AI, and my honest answer is both, often on the same day. I lead the AI task force at Xenium HR, so much of my week is spent on the exciting side. Writing is also a big part of my job, and that’s where my anxiety lives.
Jeff Wilser has the same split, and he usually admits as much in his workshops. He’s a journalist who has written eight books and reported for outlets like The New York Times and Fast Company. Through his reporting and his podcast, AI-Curious, he’s interviewed people from Sam Altman to IBM’s chief scientist. After hundreds of those conversations, he realized the questions he was asking were the same ones leadership teams were asking. That’s when he started running AI workshops for companies.
What I took from talking with him is simple. The companies that get the most out of AI bring it in with intention. They point it at growth, and they’re clear about which decisions stay with people.
Your AI Strategy Sounds Different Behind Closed Doors
Jeff has stood in front of enough leadership teams to recognize a pattern. In public, leaders say they’re embracing AI and leaning into a strategy. In private, many aren’t sure what to do or what it all means, and there may not be a clear roadmap.
“There’s so much AI hype and also AI anxiety, and a lot of companies are stuck between these two poles,” he said.
Both poles can show up on the same leadership team, even in the same department. Picture a CFO whose company policy says Copilot only but who’s building agentic workflows in Claude and ChatGPT on the side. In the same division, someone else refuses to put any data into AI at all.
Jeff’s answer to that split is curiosity. He’s heard the case for top-down rollouts and the case for bottom-up champions. In his experience, though, the surest way to get a culture on the same page is for leaders to encourage curiosity and even play, starting with low-stakes, internal uses. An edict that amounts to “thou shalt use AI” lands very differently than an invitation to explore whether AI could help with a specific problem.
When Jeff interviewed Denver Mayor Mike Johnston about getting city employees on board, the mayor described that same approach. He asks people to play with the tools for a while, then use them to find ways to solve real problems in the organization.
Leaders can also be more open about what they don’t know. “I think having a decent amount of humility can go a long way to gaining credibility and buy-in from employees,” Jeff said. In his view, leaders can be “confident in its potential without pretending to know everything.”
The AI Anxiety Gap Between Leaders and Employees
Jeff often runs a quick poll in his workshops. On the count of three, everyone calls out whether they feel more excited or anxious about AI. In a room of leaders, sometimes nearly every hand goes up for excited.
Then he changes the question. He asks those leaders to picture their most junior employees, the junior analysts and project managers, and guess how they’d answer. There’s an awkward pause, and then the answer comes back: anxious.
“There’s a gap between the excitement of leadership and the anxiety of the actual doers.”
Jeff doesn’t believe AI adoption will stick until leaders deal with that gap. He sees CEOs assume that if they’re excited, that’s all that matters. In his experience, the team’s culture and attitude are what make or break adoption.
Closing that gap starts with taking the worries seriously. Jeff doesn’t think it makes sense to tell people they’re being silly for worrying about their jobs or about cognitive offload. He also treats principled objections, such as concerns about energy and water use and how the models were trained, as fair.
It also helps to talk about AI in terms of the work people already dread. Jeff says that’s how you get the best acceptance from a team. Skip the message that AI is going to replace things, and say something closer to this:
“Hey, you know these tasks that you kind of dread doing. Good news, AI can help you with these tedious things and give you more time for stuff you actually do enjoy doing.”
Should You Use AI to Cut Jobs or Grow?
Jeff thinks one of the biggest questions that isn’t being asked enough is what a company will do with AI when the chips are down. Will it use AI to grow, or to cut?
He put that question to Eric Snowden, Adobe’s senior vice president of design, when the two shared a stage. Picture a team of four doing creative work. New AI tools mean two people can now do the same amount. Do companies cut the team to two, or keep all four and have them do more and better work?
Jeff thought it was a bit of a gotcha question. Snowden’s answer, as Jeff relayed it:
“The question is, are these companies out of ideas? If you’re out of ideas, then yeah, it’s tempting just to cut, and you shrink.”
Jeff took it from there. Cutting might help for a quarter, but over time those companies get “lapped by the companies and competitors that figured out the new ideas.”
His rough guess is that about 80 percent of companies are thinking mostly about cutting, speeding things up, and automation. He’s upfront about the fact that the number isn’t backed by any study. He also isn’t against automation. “I’m not saying automation is wrong. I’m not saying don’t automate certain tasks, but I would marry that automation with a focus on growth.”
That one hit close to home. I’ve fallen into the same trap, reaching for AI to make an existing process faster instead of using it to build something I never could have built before.
That’s why one of Jeff’s workshop modules has very little to do with AI. He asks teams what they wish they could do but can’t right now, the ideas they’ve wanted to pursue for years without the bandwidth or resources. Once people name those, they start seeing where AI could help.
He’s also noticed that size matters, again from his own experience rather than a study. Smaller teams, down to a team of one, are more likely to see AI as a way to compete and take on work they couldn’t do before. Larger companies have more at stake and tend to play defense. Jeff understands why, but he sees a long-term risk in it if scrappier teams beat them to the punch.
He points to a community bank he worked with. His website credits the bank with saving 6,000 hours of work after his workshop. Jeff is candid that the bank was already excited about AI and that he can’t take full credit. What he likes about the example is where the time went. The bank wanted to grow and knew it couldn’t do that with its current staff. As Jeff understands it, they added another branch without hiring a wave of new support staff, and no one lost a job.
What AI Theater Looks Like in the Workplace
Jeff has a name for AI adopted for its own sake. He calls it AI theater. It can look like shiny tools nobody has fit into real workflows, or getting locked into one vendor while the tools change month to month. The most egregious example he cites is token quotas, in which companies threaten to punish people who don’t use enough AI in a given week.
His favorite example came from a Colorado Avalanche game. During intermission, the jumbotron ran an AI-themed promotion that turned fans’ heads into something like dragons. To Jeff, it looked like a cheap Snapchat filter. He could picture the conference room where someone decided that hockey fans needed to see AI.
“When fans go to a hockey game, you know what they wanna see? Hockey.”
It’s a funny story, and the same thing happens inside companies when a tool gets rolled out because AI is the thing to do right now. Jeff’s filter is simple. “It’s just as important to know when not to use AI as when to use AI.”
How to Use AI for Learning Without Getting Dumber
The way I learn has changed. I’m not doing web courses or live workshops anymore. When I want to learn something, I ask ChatGPT or Claude how to do it and learn as I follow the steps.
That raises a fair question, and it’s one Jeff lives with. He shares research on the cognitive costs of leaning on AI, and he also runs workshops that get people using it more. For him, “so much of this comes down to intention and how we’re using AI.”
He borrowed a framework from Victor Varnado, a cartoonist for The New Yorker whom he interviewed. With AI, you can effectively have a conversation with Einstein, and you can do one of two things with it. You can ask Einstein to help you understand something hard and act as a sparring partner, the way an HR leader might work through a confusing new state regulation. Or you can say, “Einstein, do my math homework.”
Jeff frames it as a question of whether AI helps you get smarter or lets you get dumber. His answer is “yes.” It depends on how you use it. He compares AI to fire, which can keep you warm and heat your food, or burn your house down.
Used the first way, Jeff sees AI as “a curiosity multiplier.” One answer sparks the next question, and the more context the tool has about you or your company, the more the learning fits what you actually need. He’s bullish on AI for learning over the long term, even as the people who build courses and webinars figure out what it means for them. For HR teams that own learning and development, the Einstein test is a useful one to share with employees.
Why HR Keeps Getting Left Out of AI Adoption
AI doesn’t fit neatly into any single function, but HR plays a role in all of it because much of the adoption involves change management. When I asked Jeff where HR usually sits, his answer was candid. He thinks HR isn’t playing as big a role as it should, and that CEOs aren’t empowering HR as much as they could.
“AI is often thought of as a tech issue, and obviously it’s tech, but AI is also a people issue,” he said. So much of AI comes down to culture, and HR is a big part of culture. In his view, companies that don’t invite HR to the table enough are hurting themselves.
His practical idea is for HR to bring data. A simple anonymous survey can ask employees how they’re using AI and what they’re using it for. If you want to get a little cheeky, it can also ask whether people are using the sanctioned tools or ones nobody approved. Data like that could strengthen HR’s seat at the table.
What AI Won’t Replace in HR
Jeff is still a working writer. He’s collaborating on a book with an explorer and a broadcast legend, and he has drawn a hard line around it. “AI will not write one word of that book,” he said. Outside that line, he uses AI ambitiously to speed up research, organize notes, and keep track of transcripts and chapters. That frees up more of his time for the writing itself.
That line has stuck with me. Jeff thinks the same framework works in HR, by being “really sacred and uncompromising about where human judgment lies.” AI can do a ton of the grunt work and paperwork leading up to a decision. The decision itself, such as whether someone gets a raise or goes on a PIP, belongs to the person.
When I asked him to name one HR decision that should never touch an AI model, he went straight to termination. In his words, “No one should be fired on the basis of an AI recommendation.” The same goes for hiring and putting someone on a PIP. “The more sensitive it gets, the more human that should be,” he said. He was careful to note that he isn’t a lawyer. He still expects a company doing this would land in legally questionable territory, and he’d never recommend it on common-sense grounds either.
Jeff thinks human judgment matters even more as the tools get better. In most professions, the higher people rise, the less their work is about technical skill and the more it’s about judgment. He sees that as good news for HR, since human connection is becoming more important in the age of AI. As he put it, “human is in the title for a reason.”
When I asked what HR work he hopes people still do entirely themselves five years from now, he started with the basics. “How are we treating our people? How are we training our people? Who are we promoting?” Those questions, along with hard calls like layoffs and the work of setting culture, “should all be human things.” He expects less paperwork and fewer forms. More of the job will be reviewing accuracy and AI workflows, sometimes across several agent conversations at once, while people keep “making the human decisions that affect the humans in the company.”
How to Start Using AI at Work With Intention
Jeff calls empowerment his north star for AI. For a small HR team with no IT department, his first step is one he admits sounds a bit like a cop-out, though he says it almost always works.
Set aside half an hour, or an hour if you have it, on a Monday morning. Give an AI tool as much context about your role and your company as you can. Then ask it to act as your AI strategist and find the low-hanging fruit. Those are the things you could do right away without spending much money or investing in intensive training.
Some of what comes back will be obvious, like help with job descriptions or interviews. Some won’t. “It will almost unfailingly spark some ideas you hadn’t already had,” Jeff said.
I’d add one step before that prompt. Write down what you’re actually trying to accomplish, whether that’s clearing out the work your people dread or finally building something you’ve put off for years. That gives the tool something real to work with, and it keeps your people at the center of the work.
Brandon Laws is the host of Transform Your Workplace and the VP of Marketing and Product at Xenium HR, an outsourced HR and payroll partner for small and mid-sized businesses in Portland, Oregon.