Welcome to Episode 593
Melina Palmer: Hello, hello everyone. My name is Melina Palmer and I want to welcome you to The Brainy Business Podcast.
Have you heard the stat that 95% of organizations saw no measurable return on their investment in AI? This came from MIT’s Project NANDA in a report called The GenAI State of AI in Business 2025, and its findings were based on a review of hundreds of corporate AI deployments and interviews with business leaders. The question on everyone’s mind, of course, is what went wrong and how can we avoid it?
The report pointed to some very practical reasons, including teams building their own tools instead of buying proven ones, or deploying AI in the areas that get the most attention rather than the areas with the highest actual return. But underneath a lot of the choices sits something familiar to anyone who studies behavior: how people respond to a change they didn’t choose and don’t yet trust.
This matters right now because nearly every organization, whatever industry you’re in, is somewhere in the middle of figuring out its own AI rollout. And of course, we all want the numbers to improve. At the very least, you want to ensure that whatever you invest in is more likely to succeed than not. Today’s conversation should help with that.
I’m joined by Dr. Gleb Tsipursky, a longtime friend of myself and the show. Gleb has spent nearly 30 years studying how people make decisions at work, starting with watching the dot-com boom and bust up close in the late ’90s. He holds a PhD in Behavioral Science from UNC Chapel Hill, taught in the Decision Sciences Collaborative at Ohio State, and is the CEO of Disaster Avoidance Experts. His work has been featured in Harvard Business Review, Fortune, and Forbes, and the New York Times has called him “The Office Whisperer.” Pretty cool.
As you listen, I encourage you to think about yourself and your team and how they honestly feel about changes. Gleb is going to share some great insights for communicating with different archetypes as we go along, so having specific people and their tendencies in mind will be really helpful.
Really quickly, before we get into the conversation, I want to be sure you know that there are links in the show notes for my top related past episodes and books, ways to get in touch, and more. It’s all waiting for you in the app you’re listening to and at thebrainybusiness.com/593.
Understanding the Psychology of AI Adoption
Melina Palmer: Now let’s jump right in. Dr. Gleb Tsipursky, welcome back to The Brainy Business Podcast.
Dr. Gleb Tsipursky: It’s always such a pleasure to be back on with you, Melina. It’s great to be back, so thank you again for inviting me.
Melina Palmer: Yeah, I know it feels like we’re in contact a lot, but it’s been a while since we’ve done an interview, though I’ve had you on a couple of times and always enjoy having you here. For those who don’t yet know you, can you share a little bit about yourself and the work that you do?
Dr. Gleb Tsipursky: Sure, happy to. So my expertise is in the future of work. I’ve been doing this since, gosh, 1999, so it’s been over 26 years now—about 27 years. My background is in behavioral science. I look at how people behave and how they make decisions. I have a PhD in this topic from UNC Chapel Hill. I taught at Ohio State as a professor in the Decision Sciences Collaborative on this topic.
I also have a number of books that came out on this topic. I think you’ve interviewed me on Never Go With Your Gut: How Pioneering Leaders Make the Best Decisions, and The Blind Spots Between Us. Now we’re going to be talking about my new book, The Psychology of AI Adoption at Work: From Resistance to Results with Georgetown University Press. I’m very excited about that book.
That’s talking about exactly what the title says—truth in advertising. It’s really focusing on how to make sure that you are thinking about the psychology of your employees and your staff. How do you get people to adopt AI effectively? Because people have a lot of anxiety around AI and the real challenges with AI lie around adoption, not the technology. It’s the first book that really gets at the psychology of AI adoption.
Besides that, my work involves consulting with clients about AI adoption. AI adoption in the last few years has been my real focus. I work with small and medium-sized businesses as well as business units within larger corporations, helping them figure out how to get their employees to adopt AI effectively.
I run an eight-person consulting agency called Disaster Avoidance Experts, which you can check out at disasteravoidanceexperts.com. My work is regularly featured in venues like Harvard Business Review, Fortune, and Forbes, and the New York Times has called me “The Office Whisperer” for my focus on the future of work and ability to help folks make the best decisions in that area. So that’s what I do; that’s what I focus on.
Melina Palmer: Awesome. Yeah, I know we first talked about, like you said, Never Go With Your Gut, and then we had a whole conversation about return to work and those sorts of office policies, which continues to be an ongoing debate and discussion for companies.
Learning from the Dot-Com Boom and Bust
Melina Palmer: I’m curious, as you said, you’ve been working in the future of work for going on three decades now—do you know what first got you interested in that sort of future state of business, or did you just sort of fall into it? How did that come about?
Dr. Gleb Tsipursky: It was really around 1999 when I started getting into it. What really fascinated me was the situation with the dot-com boom and bust. I saw the dot-com boom when all these companies were making money like gangbusters in 1998 and 1999—companies like Pets.com, Webvan, and so on. They were the big things getting lots of praise in the Wall Street Journal, New York Times, and so on.
And then the dot-com bust in 2001 happened, and all of these companies went bust. They were getting slammed in the Wall Street Journal and the New York Times. But you know what? The leaders were the same, and their approach to business was the same. What changed was the situation.
It made me really realize that people don’t necessarily know what they’re doing in terms of the future of work. The kind of decisions that business leaders make are not necessarily the best, and they don’t really have a good grip on how to make good decisions as the situation changes in the future of work. That really made me want to study how leaders make the best decisions and how they behave in the most effective way to ensure that they adapt and don’t fall into any of the psychological traps that cause us to make bad decisions.
That’s what led me into my program of study in behavioral science, my first big best-selling book Never Go With Your Gut, and another book we talked about, Returning to the Office and Leading Hybrid and Remote Teams, my book about hybrid work and remote work. So yeah, that’s what really got me fascinated with these topics.
Melina Palmer: Yeah, and of course there’s always something new that’s the future, right? That’s the moving target, with new targets popping up and things to be looking at.
Why Top-Down AI Deployments Fail
Melina Palmer: It seems a reasonably clear through-line of what brought you into looking at AI, as that’s like all anybody’s talking about these days. What have you found as you’ve looked at the psychology of AI? What were early concerns companies were looking at, and what has your approach been to this topic for businesses?
Dr. Gleb Tsipursky: There was a recent study from MIT that showed that something like 95% of AI projects have failed to show actual return on investment from a profit and loss perspective. That was looking at their early adoptions and what was happening. When you look at the underlying reasons for this, it’s not the technology itself.
There is a big disparity between how executives use the technology and how staff use technology. Executives, when they’re using the technology, are finding that it’s wonderful. They’re getting like 10-plus hours of time savings per week. Then they try to implement it and tell their staff, “Use the technology, here you go.” And then, for some reason, the staff are not using it or not using it effectively.
Executives and middle managers are really surprised to see that. They wonder, “Why have I gotten all this benefit personally from the technology, but my team is not getting this benefit? Why am I not seeing it in the profit and loss?” What they’re failing to realize is that this top-down approach to AI adoption just doesn’t work for AI compared to other technologies.
Deterministic Tools vs. Probabilistic AI
Dr. Gleb Tsipursky: AI is not a deterministic technology. Previous typical technologies that people are used to are deterministic, meaning if you do this, then that next thing will happen. Like a CRM: you push a button, and it sends an email. Or a database: you push a button, find information, and tag them with it.
AI is very much not like that. It’s much more like a human—it’s a very flexible tool and it’s probabilistic rather than deterministic. “Deterministic” means you do this, you get that. “Probabilistic” means much more like a human: you say something, and the human might do it or might not, or they might not do it in the way that you want them to do it.
AI needs to be trained and customized to the specific workflows of each individual employee. So there’s much more challenge to adopting AI than to other technology. It really needs to be something that each employee customizes to their own workflows, rather than a top-down approach of just saying, “Hey, use AI.” That could work with some minor problems when implementing a new database or a CRM, but it doesn’t work very well with AI.
The Empathy Gap and Employee Anxiety
Dr. Gleb Tsipursky: The other huge area that we should talk about is how employees actually feel about AI. I do a lot of research on this topic. I did a number of survey groups and focus groups, and I talk about that in The Psychology of AI Adoption at Work. I have a chapter in the book specifically about how employees and middle managers feel about AI.
When you’re talking about company-wide adoption, executives are the ones pushing AI, whereas middle managers, not so much necessarily, and especially employees have a lot of hesitancy. For example, a survey from Pew showed that something like 50% of people are more anxious about AI adoption than excited about it. Only 10% of people are more excited than anxious.
Executives would find themselves among the 10%. Your staff would mostly find themselves among the 50% who are more anxious than excited. Meanwhile, 40% of people are equally excited and anxious, with more middle managers in that number.
When you are excited, you don’t realize the experience of people who are anxious. That’s a cognitive bias called the empathy gap. Dealing with the empathy gap, executives aren’t really realizing what’s going on with their employees. They’re not realizing that their employees are really resistant to AI usage; they’re anxious about it and trying not to use it, or when they’re using it, they’re using it badly.
It’s very easy to use AI badly versus using a CRM. It’s much harder to use a CRM badly, but very easy to use AI badly because, again, it needs to be customized to your workflows. If you don’t customize it, you’re going to use it badly. When employees try to use it without excitement, feeling anxious, and it doesn’t go perfectly in a deterministic way, they say, “Well, this just doesn’t work,” and go back to their old ways. So you’re not seeing any of the time savings that executives are expecting.
Going back to that MIT survey showing 95% of AI projects fail to show P&L return—well, that’s the reason. You really need to address people’s psychology when addressing AI, not so much technology, because the technology is great. Technology is great; psychology, not so much.
Melina Palmer: Yeah, and it’s such a key point. I remember when I first saw that MIT study—and of course people have quoted for years the not-actually-citable “80% of change initiatives fail” stat—to see another study showing an even higher failure rate was striking. But then you look at what they said, and the ones that do work have their people stuff in place. They’re really looking at the people here, which I was really pleased to see in that research. I’ve shared that far and wide with my own clients.
You have to consider this other element. I love when you were talking about that empathy gap piece. I’m sure there are a bunch of people listening who recognize that when someone is excited, it’s really hard to see those who are anxious. I think that gap is such a key point.
I would assume, too, for some of those executives rolling it out, they might have different anxieties about it, which makes it even harder to see other people’s anxieties along the way. It’s a compounding problem. Whether you layer in some of those psychological profiles you talked about or discuss that empathy gap issue, I think that’s going to really resonate with a lot of people listening.
Dr. Gleb Tsipursky: Yeah, I hope so, because they really need to hear this. They are making some bad decisions, and they need to be making better decisions about this if they want their employees to adopt it.
The Eight Archetypes of AI Adoption
Melina Palmer: Do you want to talk about some of the employee profiles that you’ve seen in your research?
Dr. Gleb Tsipursky: Sure, let’s jump there. I’ve worked with over 50 clients by the time I wrote my book, and since that time, I’ve worked with about 35 more clients, and these findings have been confirmed in those organizations as well. Through surveys and focus groups, I’ve figured out that there are eight distinct employee profiles for AI adoption.
Let’s talk about the employee profiles, from the most resistant to the most enthusiastic:
- The AI Alarmist: They are in the definitely anxious category. They’re worried AI is coming for their job—which they have reason to be, given how many announcements there are about job cuts. Their primary concerns are job loss, AI mistakes, and broader societal risks or ethical concerns. Their typical behavior is resisting AI, pointing out every error, and using AI only when forced. Their resulting work with AI is quite poor, and they expect layoffs. What they need is reassurance—a commitment to no layoffs because of AI, ensuring that employees who learn to use AI effectively will have secure jobs. They need transparent communication and guidance on how to use AI with human oversight.
- The Pragmatic Resistor: They are somewhat worried about their jobs, but more worried about change management issues and AI creating more work. They worry about AI making mistakes and replacing existing processes that already work. What helps them over the hump is evidence that AI works in their context—clear ROI from internal pilots and demonstrations using their specific department’s workflows during training, rather than generic examples.
- The Skeptical Observer: They are willing to try it out, but they want more proof. Their mindset is, “I’m not convinced, but I’m watching and open to it.” They are moved by watching their peers experiment and succeed. They need good metrics, peer examples, and training context centered around what their coworkers are doing.
- The Apathetic Bystander: They perceive AI as not really affecting them. They feel, “Let me do my thing, and other people can adopt AI.” They don’t show much resistance, but they don’t show engagement either. They will ignore optional training, so training must be required for everyone. They need clear personal relevance and to see that their direct supervisor expects them to use it.
- The Reluctant Adopter: They use AI because they have to under pressure from leadership. They worry about falling behind, so they follow instructions carefully but don’t explore independently. They need confidence building, mentoring, and clear success metrics.
- The Cautious Optimist: They believe AI looks promising, but only if used responsibly. They want clear guardrails and practical tips. They are willing to learn and take initiative within a framework of good governance and incremental wins.
- The Efficiency Seeker: These are the optimization hackers who try to be as pragmatic as possible. They want to eliminate repetitive work and boost output. They adopt AI quickly and look for new use cases independently. They don’t need AI 101; they need advanced optimization training. They are often already using “shadow AI” on their personal devices. They respond well to recognition of their productivity.
- The AI Evangelist: They see AI as truly transformative. They are innovative, want to learn continuously, and want to help others adopt AI. They experiment, teach coworkers, push others to use AI, and discover organizational use cases. They need guidance on how to train others constructively, along with governance boundaries so they don’t accidentally leak company data using unapproved tools. They make ideal members for AI committees and peer-mentoring programs.
Melina Palmer: Thank you! That is really helpful to have them broken out in that way. I’m sure everybody listening is mapping people on their teams to those types.
Strategies for Moving Teams from Resistance to Results
Melina Palmer: If an organization has all eight of those profiles, people aren’t going to self-select or opt into “training for alarmists.” What is your recommendation for leaders bringing everyone along without triggering alarmists or boring the evangelists?
Dr. Gleb Tsipursky: The most fundamental thing you want to do is figure out how to get buy-in from the resistors, because they are the ones who will slow down your team and undermine ROI.
I find it really important to run anonymous surveys and focus groups to uncover their actual concerns. You kill two birds with one stone: first, you get accurate information about what is causing their worry; second, you make them feel heard. Of course, you must take visible action based on the feedback.
Surveys and focus groups provide the basis for town halls and follow-up communications. For instance, addressing job security directly or reassuring employees that learning AI will protect their roles.
For people who say they are too busy putting out fires to learn AI, show them concrete examples of how Efficiency Seekers and AI Evangelists have saved time. It is very realistic for an office worker to save over 10 hours a week within the first couple of months of proper training.
When I conduct half-day workshops, employees don’t just listen to lectures—they spend time building custom prompts, agents, and automations for their actual workflows. They leave the workshop with immediate tools that save them time right away.
For example, building a tool that summarizes incoming emails and drafts automatic replies, or an automated meeting preparation tool that pulls relevant documents across Google Drive, SharePoint, and your CRM to synthesize a brief. Cutting prep time from 15 minutes down to 3 minutes per meeting creates immediate buy-in. When people build tools for their own needs, they move from resistance to engagement.
At the same time, you must establish official AI governance and privacy policies to reassure those concerned about data leaks, while giving Evangelists approved sandboxes so they don’t resort to risky shadow AI.
Melina Palmer: Those are great examples. Nobody likes spending time manually gathering prep documents or sifting through endless emails. Giving people agency to build what helps them most allows them to safely test AI and build momentum.
Case Study: Empowering Employees vs. Replacing Them
Melina Palmer: We often see scary headlines about AI causing layoffs or projects failing. Do you have an example of a company that got AI adoption right?
Dr. Gleb Tsipursky: Absolutely. I recently worked with a CPA firm that was considering two paths. Option one was buying an expensive AI agent designed to automate junior CPA functions, effectively replacing staff. Option two was doing an adoption training focused on empowering their existing team.
They chose the training approach. Instead of replacing people with AI, we trained the staff to become managers of AI tools and agents.
For instance, they built an AI tool connected to client data in QuickBooks that runs weekly, updates records, and drafts a customized financial report. Previously, running that report took a substantial amount of manual time. Now, the report runs automatically, freeing up the CPAs to analyze the data, identify cost-saving strategies, and provide high-level advisory services to clients.
The result? The firm retained clients better, expanded its service offerings, and grew revenue—all without laying anyone off. The staff shifted away from repetitive manual tasks toward high-value client relationship management that AI cannot perform. That is where real ROI comes from.
Melina Palmer: That is a fantastic outcome and a clear roadmap for organizations looking to scale AI thoughtfully. For listeners who want to get your book and learn more about your work, where should they go?
Dr. Gleb Tsipursky: You can find The Psychology of AI Adoption at Work: From Resistance to Results on Amazon, Barnes & Noble, or directly through Georgetown University Press. You can also visit my website at disasteravoidanceexperts.com/aibook to find the book, along with bonus assessment tools to evaluate your organization’s AI readiness.
Melina Palmer: Perfect. We will have those links in the show notes. Thanks again, Gleb, for joining me today—it’s always a delight!
Dr. Gleb Tsipursky: It’s been great! Thank you again for having me, Melina.
Key Takeaways and Wrap-Up
Melina Palmer: Thank you again to Dr. Gleb Tsipursky for joining me on the show today. What got your brain buzzing in today’s conversation?
For me, while I love understanding how people are the same at their core, I also think it’s so important to remember where we’re different. That comes out a lot when you look at segmentation. In marketing, segmenting audiences is standard practice; you wouldn’t send the exact same message to a loyal customer as you would to a brand-new prospect.
Yet internally, organizations often throw those best practices out the window when rolling out new tools like AI. They send one blanket email, host one mandatory training, set one deadline, and move on.
That is why I really appreciate the eight profiles Gleb shared today. Understanding whether you are speaking to an AI Alarmist, an Efficiency Seeker, or a Cautious Optimist allows you to tailor your communication. What sounds exciting to an Efficiency Seeker might sound threatening to an Alarmist. Thoughtful segmentation makes a massive difference in adoption—especially when 95% of AI initiatives struggle to show a return due to human resistance.
Which of these profiles resonated most with your team? What will you do differently during your next rollout? I’d love to hear your thoughts! Connect with me on social media as @TheBrainyBiz or find me directly as Melina Palmer on LinkedIn.
Don’t forget to check the show notes in your app or at thebrainybusiness.com/593 for links to Gleb’s book, related past episodes, and additional resources.
Join me next time for another insightful episode of The Brainy Business Podcast. Until then, thanks for listening, and remember to be thoughtful!
Announcer: Thank you for listening to The Brainy Business Podcast. Melina offers virtual strategy sessions, workshops, and other services to help businesses be more brain-friendly. For more free resources, visit thebrainybusiness.com.
