Resistance is a spec, not a symptom

The people in companies fighting against AI are the most valuable asset right now.

I used to think resistance was the enemy of progress.

A manager blocks AI-generated reports because a single hallucination in a client deck could end the relationship.

A logistics coordinator refuses to train the routing algorithm because the last team that automated their workflow lost four people in six months.

An underwriter refuses automated risk scoring because twenty years of pattern recognition shouldn’t be reduced to a checkbox.

For years, my instinct was: get them on board. Sell them the benefits. Convince them.

Then I started paying attention to what they were actually protecting.

Corporate antibodies

Corporate antibodies seek and destroy innovative projects.

You’ve spent months building something that works. The pilot works. The value is there. You’re ready to scale, and bring your innovation into the organization.

That’s when the antibodies activate.

• “We can’t do it, the roadmap is full.”
• “We can’t do it, legal needs to review this.”
• “We can’t do it, this cannibalizes our existing product.”
• “We can’t do it, my team wasn’t consulted.”
• “We can’t do it, we tried this before. It failed.”

All reasonable. All lethal.

The antibodies are not the enemy. Their job is to protect what exists today. Protect what was built. Protect the revenue of today.

Change doesn’t just add value. It displaces existing roles, status, and safety nets.

These aren’t objections. They’re specs.

Gartner lists five fears behind employees’ resistance to AI: losing their job, or seeing it become harder and less interesting; AI that produces incorrect or unfair insights; not knowing where, when and how the company uses AI; reputational damage when AI is used irresponsibly; personal data put at risk.

Every one of these is something a well-designed deployment should account for.

These aren’t objections. They’re specs.

Addressing them isn’t slowing the rollout. It’s designing it for the organization.

How Pernod Ricard reached 85% adoption

Take Pernod Ricard.

The spirits group built D-STAR, a machine learning system that recommends which stores its sales reps should visit and which products to push. In Germany, reps worried it would diminish the strategic side of their job, the part many valued most.

Instead of forcing adoption, leadership ran A/B tests to prove the tool worked. In France, stores where reps followed the recommendations saw better market share and sales growth than the control groups.

Reps who followed the recommendations but missed their targets were not penalized.

Every market rollout had its own deployment team: local change management specialists, data analysts, trainers.

By 2023, D-STAR had reached 85% adoption across the markets where it was deployed.

Not because resistance was overridden, but because it was addressed.

Listen to what they protect

You don’t fight antibodies. Listen to what they’re protecting. Understand why they were triggered. Address what triggered them.

The next time someone pushes back on something new, try something different.

Don’t ask how to get them on board.

Ask what they’re seeing that you’re not.

What would change if you treated resistance as a spec, not a symptom?

Sources: Gartner, Overcoming Employee Fears of AI to Drive Business Value (June 2024); Harvard Business School Working Knowledge, How Pernod Ricard Stirred Up Employee Enthusiasm for AI (November 2025), on the case “Pernod Ricard: Uncorking Digital Transformation” by Iavor Bojinov and Edward McFowland III (2024).