🎧 Listen: Castos | 📺 Watch: full episode on YouTube
This week on the Unscripted Small Business Podcast, host Jeremy Rivera sits down with Dr. Gleb Tsipursky, CEO of Disaster Avoidance Experts and author of The Psychology of AI Adoption at Work: From Resistance to Results (Georgetown University Press). If you’ve built something clever with AI, shown it to your team and watched the room go quiet, this episode explains the silence. Gleb’s short version: the technology isn’t what’s failing. Your employees are reacting the way any rational person would.
Read more from Gleb on the Disaster Avoidance Experts blog.
Owners are excited about AI, and employees are anxious
Gleb opened with the research he cites in the book. According to the MIT research he points to, 95% of AI initiatives fail to scale when a business tries to roll them out. He also cites a Boston Consulting Group study putting the share that show value across the business at around 6%. And he points to Pew Research: in Pew’s June 2026 survey, 52% of Americans said they’re more concerned than excited about AI in daily life, and only 9% said they’re more excited than concerned.
You can probably guess which camp the owner is in. Gleb’s point is that the owner’s camp is the problem.
“So there’s a fundamental empathy gap, emotional gap, between the intuitions of what leaders feel about AI tools and what their employees feel about AI tools.”
That’s the first question he works through with every small business client, and it’s why he wrote the book. In his words:
“The leader needs to understand that her or his gut does not lead to the right outcome with AI adoption because their gut emotions, their intuitions are at odds with the emotions of their employees.”
It’s group dynamics, not the technology
Jeremy asked whether AI is just overhyped, or whether the rollout itself is where things break. Gleb didn’t hesitate: “No question, it’s group dynamics.” He cites McKinsey research finding that 80% of respondents get individual productivity from AI, while only around 6% of businesses see a material impact on the bottom line.
“AI is showing individual productivity, but the group dynamics is where it’s at. And that’s not the technology.”
He framed people as “biological intelligence,” trained on past patterns the way AI is. So when the owner shows off a demo, what does the team actually hear? Gleb runs an eight-person firm himself and admitted he has to hold back “so they’re not afraid of what will happen.”
“When you show them this cool demo, you show them this really cool thing that you built, and you’re so excited”
You see a business moving forward. Your team, he said, is thinking: “This thing will replace me.”
AI workslop is what resistance looks like
When people are forced to use a tool they’re afraid of, they don’t refuse. They comply badly. Jeremy’s friend Michael McDougald of Right Thing has done research on how to identify AI slop, and Gleb’s point is the human side of the same problem: why your own team would produce it.
Gleb calls it AI workslop, citing a Stanford and BetterUp Labs study that Harvard Business Review published. As he describes it, 40% of respondents had received workslop within the last month, and each instance took about two hours to correct.
“And that’s poor AI outputs that look polished, but they have some problems in them.”
A report without the right data. A PowerPoint without the right content. “It’s AI being used badly,” Gleb said, and it shows up when the owner pushes the tool on people who are “worried, scared, anxious about them.”
Why AI isn’t another CRM rollout
Jeremy brought up his conversation with Keith Perhac of SegMetrics, who described employees in Japan treating AI as a shared productivity project instead of a threat. Gleb works mainly across the US, Canada and the UK and hasn’t worked inside Japanese companies, but he said he has seen international surveys showing that the US is one of the most anxious countries for AI adoption.
Then he explained why that anxiety hits AI harder than other software. With a CRM, you write an SOP, train everyone the same way, and check the work. AI doesn’t work like that.
“You can stand over somebody’s shoulder and see if they’re using a CRM correctly, because that’s something you can measure.”
With AI, each person has to fit the tool to their own workflow. “It’s kind of like hiring an executive assistant for yourself,” he said. You can’t run an IT-led implementation for that, so you need people motivated to experiment, and fear works directly against that.
The three people quietly resisting your AI rollout
Gleb says the book draws on over 100 consulting projects, 600 focus group responses, over 10,000 survey responses and more than 50 interviews with leaders from companies like Microsoft, Experian and Uber. From that he sorts the resistance into three psychographic profiles:
- AI alarmists. The biggest group. “They’re basically afraid that by using AI, they’re training their own replacement.”
- Pragmatic resistors. Usually more senior. Their jobs feel safe, but their identity doesn’t: creatives who write copy, finance pros who build the analysis. “What they’re worried about is identity threat.”
- Reluctant adopters. They use AI already, and keep it quiet because of the stigma. “And so they use AI tools, but they don’t talk about them.”
What to do about each one
Chapters seven through nine of the book cover the fixes. Here’s the version Gleb gave you on the show.
For the alarmists, commit to growth, not job cuts, and flip the script. Gleb points to a Stanford study he says found that AI-forward companies grow top-line revenue over 9% faster than less AI-forward peers in the same industry, and grow headcount 6% faster. Then tell people plainly who stays:
“It’s not by using AI tools that you’ll be replaced. It’s that if you don’t learn how to use AI effectively, you won’t really stick around here.”
He adds that the skill travels with them. He cites a PwC survey finding that people with domain expertise plus AI skills command a 69% wage premium.
For the pragmatic resistors, survey the tasks they hate. Don’t start with the work they’re proud of. Ask what they want off their plate, and what blocks the work they love. Then build prompts and agent skeletons around those answers. Gleb called it “you go slow to go fast.”
For the reluctant adopters, go first. Talk openly about how you use AI for your emails, board decks, sales outreach and performance reviews. Open a Slack channel, Teams channel or Trello card for AI use cases, and praise what people share. His small-business version of a reward:
“I would recommend doing something like a $25 Starbucks card every month for the person who shares the coolest AI use case”
Resistance is the rational move from your employee’s chair
Jeremy asked how Gleb handled this in his own firm. He admitted he started out “doing too much of the demoing of cool AI use cases” until he watched how his employees reacted. His first AI book, ChatGPT for Leaders and Content Creators, came out around January 2023 and resonated with owners and solopreneurs. His own team’s reaction was different: “this is going to replace me.” That’s where the new book came from.
Near the end, Jeremy described the gap between his own entrepreneurial wiring and people who prefer predictability and a paycheck. Gleb said resisting AI is the logical choice when you believe the tool is training your replacement: malicious compliance, workslop, then “I’m not gonna be using AI anymore.”
“Why would you use something that by using it you are going to be replaced, right?”
Gleb doesn’t think that fear is paranoid. He said that “in some companies it’s exactly what’s happening,” and he named Meta. Here’s what the reporting shows. In April 2026, Reuters reported that Meta was installing software to capture U.S. employees’ mouse movements, clicks and keystrokes as training data for AI agents meant to do work tasks autonomously. Meta said the data would be used only for model training, not performance reviews. In August, a Reuters investigation into Meta’s AI restructuring plan reported internal planning for AI to take over much of the daily work of thousands of employees, and that many staff believed they might be training their own replacements. Meta confirmed the project but said its scenarios included both layoffs and redeployments. So when your team worries about the same thing, they’re not making it up.
His fix is the same one he started with. Focus the company on growth, say so, and mean it. Otherwise, as he put it, “if you were in their shoes, totally, totally reasonable thing to do.”
Before your next AI demo, ask what your team hears when you show it. Then do three things this month: tell people in writing that you’re using AI to grow, not to cut; ask each person which tasks they’d most like off their plate; and share one way you used AI yourself this week. For more owners working through the same questions, browse our AI for small business hub.
Also on Unscripted Small Business
Gleb’s AI adoption advice pairs with Richard Hickman’s on the outward-facing side: how to build a website for AI search.
Two more takes on this conversation: what it means for marketing teams that quietly resist AI tools on SEO Arcade, and Jeremy’s own view as an owner who’s in the excited camp.
Connect with Gleb
- The book: The Psychology of AI Adoption at Work (also at Georgetown University Press)
- Website: Disaster Avoidance Experts
- Blog: disasteravoidanceexperts.com/blog
- About Gleb: Dr. Gleb Tsipursky
- LinkedIn: Dr. Gleb Tsipursky
Want the step-by-step version? Use the free AI adoption SOP built from this episode – eight steps and a printable team worksheet.
🎧 Listen: Castos | 📺 Watch: full episode on YouTube
Keep going → Browse the interviews