Overview

You announce AI adoption. Your engineers respond with eye rolls. "Not another fad." Or: "This is going to make my job irrelevant." Or worse: "We don't need AI. We need to hire good engineers and let them do their job."

This is the resistance. It's real. It's legitimate. And if you don't handle it well, it kills your transformation before it starts.

The mistake leaders make: trying to argue people into agreement. Sending emails about AI importance. Forcing adoption policies. None of that works. What works is showing people directly: through their own experience, with their own problems, how AI actually helps.

This is about turning skeptics into advocates. Not through rhetoric, but through evidence.

Understanding the Types of Resistance

Type 1: The Skeptic

"This is a fad. I've seen overhyped technologies before. AI will be forgotten in five years."

Root cause: They've experienced failed technology adoption before. They're protecting their time and energy from what feels like another distraction.

How to address: Show data. Not hype, data. "Code generation is 10% faster with AI assistance. Here's the study. Here's our internal data." Let them try it themselves on a real problem. "Spend one hour using this. See what you think." Most skeptics who actually try good AI tools change their minds.

Type 2: The Threatened**

"This will replace me. My job is going to disappear."

Root cause: Legitimate fear. AI is genuinely changing what it means to be an engineer. They're worried about job security.

How to address: Be honest. "AI will change the job. It won't eliminate the job." Show concretely what changes. "You'll spend less time typing code. You'll spend more time thinking about problems, designing better systems, mentoring others. Those are the parts of engineering that are actually valuable." If they see that their job is changing but valuable, the fear decreases.

Type 3: The Pragmatist**

"This is nice, but we have real work to do. We don't have time to learn new tools."

Root cause: They're busy. They have sprints to finish, features to ship, bugs to fix. AI feels like overhead.

How to address: Prove it saves time, not adds to it. Don't ask them to spend weeks learning. Show them: "15 minutes to set up. Here's a specific task you do all the time. Here's how to use this tool for that task." If they see 20% time savings on something they actually do, they'll adopt.

Type 4: The Ideologist**

"We should only use open-source AI. Or only proprietary. Or we shouldn't use AI at all because it's unethical."

Root cause: Strong beliefs about how technology should be done, or concerns about ethics and safety.

How to address: Respect the belief. "That's a real concern. Here's how we're thinking about it." You might accommodate them (allow open-source alternatives, or provide opt-out paths). You won't convince them, but you can make space for their values while still moving forward.

The Resistance Principle: You can't argue someone into adoption. You can provide experience. Experience beats argument every time. Get skeptics to try AI on a real problem. Let them discover the value themselves.

Turning Skeptics into Advocates

Step 1: Pick the Right Skeptic**

Not all skeptics are equal. Some are thoughtful and could be convinced with evidence. Others are ideologically opposed. Start with the thoughtful skeptics. If you convert them, they're powerful advocates because they have credibility (they were skeptical, now they're convinced).

Step 2: Identify Their Real Problem**

Don't try to convince them AI is important in general. Identify a specific, painful problem they face. Code reviews take too long? Boilerplate code is tedious? Debugging is painful? Start there.

Step 3: Show AI Solving That Problem**

"You spend 4 hours a week on code review. Claude can do a first pass in 2 minutes. This saves you 3 hours a week. Here, try it on this PR. See what you think." This is not an argument. It's evidence.

Step 4: Make It Easy for Them to Win**

They try it and it works. They feel smart. They feel like they discovered this, not that you forced it on them. This is powerful. They become advocates because they own the success.

Step 5: Have Them Share the Win**

"You found a cool way to use Claude for code review. Present that to the team. Show them what you learned." Now they're advocates, sharing their experience with peers.

Addressing Specific Concerns Head-On

"AI will replace engineers."**

Honest answer: Some tasks might be automated. Boilerplate code generation, documentation, routine code changes. These tasks are being automated. But they're not the valuable part of engineering. The valuable part is: understanding problems, designing solutions, mentoring others, making trade-offs, thinking about long-term architecture.

If you're only doing boilerplate code work, you should be concerned. But good engineers do more than that. AI is a threat to mediocre work, not to good engineering.

"The AI outputs are low quality."**

This is sometimes true. Show them the limitations. "Claude is great at generating code for common patterns. It's weak at complex reasoning or novel problems. Use it where it's strong. For the hard stuff, you're still needed."

Better: show them how to use it better. "Most people prompt Claude poorly. Here's how to write prompts that get good results." Often the issue is not the AI, it's how people are using it.

"We'll become dependent on AI and lose skills."**

Legitimate concern. If someone uses autocomplete for everything, they might not learn how to code. But the same is true of other tools (IDEs, linters, libraries). The answer: intentional learning and using AI strategically.

"Use AI for routine work. Use your brain for learning and complex work." As long as you're being intentional, you don't lose skills.

"What about data privacy? Will our code go to vendors?"**

This is important. Be clear about your policy. If you have an internal AI platform, code stays internal. If you use vendor APIs (Claude, GPT), code goes to vendors. Some organizations need on-premise solutions. Acknowledge the concern and provide options.

Addressing Concerns Authentically: Don't dismiss concerns as "unfounded" or "irrational." Take them seriously. Some are legitimate. Your job is not to convince people away from legitimate concerns. Your job is to address them directly or provide alternatives.

Case Study: Converting Skeptics Through Real Results

A mid-market fintech company had a 40% skepticism rate about AI among their 60 engineers. The CTO didn't push adoption. Instead, she identified 3 vocal skeptics and talked to each: "What's your actual concern?" One was worried about job security. One thought AI outputs were low quality. One was concerned about data privacy (financial data).

Rather than arguing, the CTO addressed each concern directly: 1) Job security: "Show me your last 5 code reviews. How much was architecture/design vs. boilerplate feedback? That's what I need your brain for. AI can handle boilerplate." The engineer realized 60% of his review comments were style/formatting, AI could handle that, freeing him for architecture work. 2) Quality: "Try Claude on your next code review. Score the suggestions. If > 50% are useful, you keep using it. If < 50%, you stop and tell others it doesn't work." The engineer tried it, found 70% of suggestions were useful, became an advocate. 3) Privacy: "Our financial data is sensitive. We'll use an on-premise AI instance. Data never leaves our systems."

Within 3 months, 65% of the team was using AI regularly (up from 0%). Skeptics became early adopters. Why? Because their actual concerns were addressed, not dismissed.

Building Adoption Momentum

Early Wins Matter Most**

The first month of AI adoption in your org determines whether it succeeds or fails. If people have good experiences early, they adopt. If they have bad experiences (AI tool breaks, output is garbage, integration is hard), they give up and tell others it doesn't work.

Over-invest in early experience. Make sure tools work. Make sure people are trained. Make sure they have support. This is not the time to be lean.

Celebrate the Skeptics Who Convert**

"Sarah was skeptical about AI. She's now saved 20% on code review time using Claude. Here's what she said." These stories are powerful. Peer adoption is driven by peer stories, not by leadership mandates.

Create Safe Spaces to Experiment**

People are scared of looking foolish. Create psychological safety. "Try this tool on a small task. No judgment. Report back." If people know they won't be mocked for trying and failing, more people try.

Find and Empower Champions**

In every team, there are 1-2 people who will naturally adopt early. Give them visibility and support. They become champions who help their peers. "Sarah's our AI expert. She can help you get started." This scales much better than centralized training.

Common Mistakes in Managing Resistance

Being Preachy**

"AI is the future." "You have to adopt this." "Stop resisting." This doesn't work. People push back harder when they feel forced. Instead, provide evidence and let them decide.

Dismissing Legitimate Concerns**

"That's a concern, but don't worry about it." This erodes trust. Either address the concern seriously or acknowledge that it's real and you're working on a solution.

Forcing Adoption Before Tools Are Ready**

You release an AI tool that's half-baked. People try it, it doesn't work well, they give up. Now you've lost credibility and made them more resistant. Better to spend extra time getting the first version right.

Only Talking About Benefits, Not Challenges**

"AI will save you time and make you better." But you also need to address: "It changes your workflow. You'll need to learn new skills. Some tasks will feel different." Being honest about challenges builds credibility.

What to Do Monday Morning

• Identify your top 3 skeptics: who are the influential people in your org who are most doubtful about AI?
  - Understand their specific concern: it's not "AI in general," it's something specific. What is it?
  - Find a way AI solves that specific concern: What problem do they have that AI is actually good at?
  - Have a conversation: "I know you're skeptical about AI. I respect that. But I think there's a way it could help with [specific problem]. Interested in trying?"
  - Make it easy to try: Don't ask for weeks of learning. 30 minutes. One real problem. See what happens.
  - Support them: If they try it, be available to help. Remove friction.
  - Celebrate if it works: "Hey team, Sarah tried this approach with Claude and saved 3 hours this week."

FAQ

Q: What if someone absolutely refuses to use AI?**

A: That's their choice. Don't force it. Make it clear that it's part of how your org works, but provide alternatives. Some people will come around with time. Some won't. Accept that. Your job is to enable the majority, not convince 100%.

Q: How long before skeptics typically convert?**

A: Weeks to months if you do it well. If someone tries AI on a real problem and sees results, they usually convert within a month. If you try to convince them through argument, they might never convert.

Q: What if skeptics are right? What if AI isn't actually useful for our team?**

A: That's possible. Some domains benefit from AI more than others. If you genuinely try (good tools, good support, time to learn) and adoption doesn't happen, maybe AI just isn't valuable for that team's work. That's ok. Be honest about it and move on.

Q: Should we make AI adoption mandatory?**

A: Mandates create resentment. Better: create incentives (recognition, time savings, interesting work) and support (tools, training, champions). Most people will adopt when they see value. Don't force the holdouts.

Q: How do we keep skeptics from poisoning the well?**

A: One vocal skeptic can undermine adoption for others. Best approach: address them directly and transparently. "I hear your concerns. Let me address them." If you handle them well, they either convert or at least they're less influential. Either way, momentum continues.

Q: What if someone had a bad experience with AI tools in the past?**

A: Bad experiences create lasting skepticism. "I tried chatbots and they're useless" means they've now decided AI is useless in general. Don't dismiss this. Ask what their bad experience was. Show them how Claude/modern AI is different. Start small: "This is a different tool. Try it on a specific task." They might be surprised.

When This Goes Wrong: Forced Adoption Without Support

A company announced AI adoption as mandatory. They set up tools but provided minimal training. Engineers tried the tools, had bad experiences (slow, confusing, low-quality output), and gave up. They now actively avoid AI and discourage others. The forced adoption backfired, creating resistance that will take years to overcome. Lesson: support must exceed mandate. If you're forcing adoption, you must make it work for people or you'll create lasting damage.

Resistance to AI is real and mostly legitimate. Don't try to argue people into adoption. Give them direct experience with AI solving a real problem they care about. Let them discover the value. This turns skeptics into advocates. The key: start with thoughtful skeptics, identify their real concerns, show AI solving those specific concerns, make it easy to try, celebrate wins. Over time, resistance converts to momentum.

On This Page

Watch the Lecture
Types of Resistance
Turning Skeptics into Advocates
Addressing Specific Concerns
Building Adoption Momentum
Common Mistakes
Monday Morning Action
FAQ

Chapter Details

Part ofChapter 7