When AI feels like a threat, not a tool

The C-suite's euphoria over AI is bumping into a harsh reality: workers who are nervous about being replaced, sidelined, or stripped of meaning in their careers.

To some employees, AI feels like the brilliant intern who is eyeing their job. 

While executives often see productivity gains in AI efficiency, workers may fear they’ll be replaced, especially those in lower-level positions. This disconnect between leadership and employees can slow AI adoption.

"We've done a disservice by anthropomorphizing AI—talking about agents as employees or coworkers," says Matt Kropp, managing director and senior partner at the Boston Consulting Group, who spends his days helping companies carefully plan the best way to implement AI in their organizations. "This creates the mental framework of AI as a threat rather than a tool."

Kropp isn't your typical ivory-tower market analyst. He's battle-tested. Before joining the firm, he spent nine years building and running tech consulting firms, including a medical informatics software company and a digital freight marketplace.

We recently spoke with Kropp about misconceptions surrounding AI adoption, including why workers resist, what leaders get wrong, and his five-stage roadmap for successful AI implementation.

Why are employees resistant to AI adoption?

First, there is skepticism about its capability. People don't believe AI does their job as well as they can. They may have tried it once, got a bad result, which can happen with probabilistic models, and concluded that this thing is no good.

Then there are ingrained habits. People are reluctant to change how they work, even if they see potential benefits. It's like personal fitness. Knowing how to use exercise equipment isn't enough; you have to create the habit of going to the gym.

Read also: Researcher Matt Beane on how AI can help to solve a talent problem it helped to create 

The most significant resistance comes from seeing AI as a threat. That’s the hardest barrier to overcome. Someone may think, "I'm an engineer, I write code—that's what makes me special. If AI can write code, why am I special anymore? Who am I?" 

What are the most effective approaches for getting employees excited about AI tools?

Training alone isn't enough. To overcome all three barriers, you need experiential learning combined with work redesign. We call this "find the toil and find the joy."

You must work with employees to jointly redesign their processes, consciously removing the toil (stuff they don't like) while preserving the joy (what they do like), so they have a better job in the end. Peer-to-peer coaching is particularly powerful.

We recently trained all our engineers on agentic coding through a hackathon format. After classroom training, we gave them seven hours of tasks impossible to complete without AI, ending with a "shark tank" presentation with prizes and recognition. Throughout the event, I watched people transition from "I don't think this stuff works" to "Oh my God, it works, and I can see how it works for my work."

The most powerful moments happen when people sit with peers, try prompting, get bad results, and their peer says, "Your prompt isn't very good—you didn't give enough context. Here's how I'd do it." That shoulder-to-shoulder sharing, social pressure, and peer learning help people let their barriers down.

 Will the increased use of AI impact employee burnout and job satisfaction?

Absolutely. At first, people fear losing something that makes them special, but once they get there, they realize they lost something like typing, which they didn't enjoy anyway. At the same time, they gained the ability to create more, which they love much more than the typing they lost. Everyone who makes it through the transition to AI is overwhelmingly positive. Engineers who become agentic coders, for instance, all say, "I'm never going back. I love working this way. I'm so much more powerful, creative, and productive."

What are some ways that leaders can lessen worker angst about AI and ensure that it’s more enthusiastically received in the workplace? 

Let's start with middle management. Managers must use the tools themselves, visibly. Teams where managers use AI are four times more likely to adopt than teams where managers don't. Managers also shouldn't sandbag—if your boss asks for AI efficiency and you claim it'll only work 5%, you're protecting your budget but hindering adoption.

At the senior leadership level, the tone you set about AI objectives is crucial. Saying "we want 30% headcount reduction through AI" won't get willing adherence. You might still make it happen through brute force, but that's not what an organization needs. A better framing would be "This is about making us more successful, innovating, moving faster, satisfying customers, and making better products."

What changes should employees expect? How does AI integration affect workflow management and workplace dynamics?

Roles start to collapse. You may need more senior people that are more able to think about, you know, the architecture, the implications, and the business objectives. A product team may transform to become smaller, more senior, more agile, move faster, and have a different process. 

Timelines will also change. Let’s focus on coding. Traditional software development involves sequential handoffs: the product manager writes specs in weeks, the designer creates mockups in days or weeks, and the engineer architects in weeks. With agentic tools, the product manager can vibe-code a working prototype in an hour, show it to users, get feedback, and iterate with designers and architects who now understand the vision much better.

What about emotional support for workers facing these changes? What should a company offer?

That's an interesting question. Executive teams understand the psychological barriers to adoption, but I haven't seen anyone thinking about providing psychological support around AI adoption. It will become an issue as people who've done their jobs for 20-30 years suddenly see their future looking cloudy. That's an issue we need to address.