The Four Faces of AI Resistance
A field guide for organizations navigating the gap between AI adoption and AI execution.
You already know AI works. The question is why it is not working consistently, and who is standing between the rollout and the result. Resistance is not the obstacle. Handled well, it is the fastest path to execution.
Developed by Eve Kedar, EdD. Delivered as a live working session for the AI Leadership Institute community, July 2026.
What it sounds like: "My way already works." "AI can't do what I do." Points out every flaw in the output, tries nothing twice.
What is really happening: Twenty years of hard-won craft feels discounted overnight. Pride is protecting identity, not resisting software.
Recognition test: Pride resists publicly. If they argue in the meeting, you have Pride, not Fear.
The play: Draft your skeptics. Make them the expert of the rollout. Their craft becomes the quality bar the tool has to clear.
The Premise
The Four Faces
I. Fear: “Will this replace me?”
What it sounds like: "I just haven't had time to try it." Quiet non-usage, no argument and no adoption. Enthusiasm in the meeting, zero logins after.
What is really happening: The tool reads as a threat to their value, not an addition to it. Fear never announces itself as fear. It hides as busyness.
Recognition test: Ask them to show you one thing they tried. Fear has nothing to show and a reason ready.
The play: Shrink the stakes. One small, private win in their own workflow. No audience, no scoreboard, no announcement email. Fear dissolves through competence, not messaging.
II. Pride: “I’m the expert here.”
Adoption is a Login. Execution is a behavior
AI is everywhere. Impact is not. Only one of those shows up in revenue.
Stop asking why people are resisting. Ask what kind of resistance you are looking at, then respond in kind. The four faces look identical from the corner office and require four different responses.
III. Scar Tissue: “We’ve tried this before.”
What it sounds like: "Remember the CRM rollout?" "This is this year's shiny thing." Compliance without belief. Attends training, changes nothing.
What is really happening: They watched three initiatives launch, stall, and quietly die. Their skepticism is pattern recognition, and it was earned.
Recognition test: Ask what would have to be different this time. Scar Tissue has a precise, historical answer.
The play: Honor the graveyard. Name the failed initiatives out loud, then show what is structurally different. Measure execution, not attendance.
IV. Discernment: “This output is wrong.”
What it sounds like: "The numbers in this draft are invented." "It misses what the customer actually said." Uses the tool, documents where it fails.
What is really happening: This is not resistance. This is quality control wearing resistance's clothing. Crush it and you lose your best safeguard.
Recognition test: Discernment comes with receipts. Specific examples, not general objections.
The play: Give them the QA seat. Formalize their critique into the workflow. Their objections become your evaluation criteria.
Why this matters more in the agentic era
When AI moves from assistant to agent, taking action autonomously, the cost of unmanaged resistance multiplies. The discerning resister is no longer friction. They are a critical control.
Suppress dissent during adoption and you will have no immune system when agentic systems make confident errors at scale.
What never works
The mandate. Usage dashboards create logins, not behavior. People are excellent at complying invisibly.
More training. Training solves a knowledge gap. None of the four faces is a knowledge gap.
The hype campaign. Evangelism deepens Scar Tissue and insults Discernment. Your two smartest resisters check out first.
You can mandate attendance. You cannot mandate belief.
Three questions to pressure-test any AI initiative
1. Show me the workflow. What changed in how work actually happens?
2. Show me the resister. Who challenged the approach, and what did they teach you?
3. Show me the number. What business metric improved beyond usage rates?
If a leader cannot answer all three, the initiative is reporting activity, not producing results.
The Playbook
The Leader’s Playbook
Draft your skeptics. Turn critics into reviewers and red-teamers.
Design for the workflow. Design around the moment where work happens, not around the tool.
Measure execution. Track cycle time, quality, error rates, and win rates, not logins.
Build the high-fence playground. Psychological safety plus visible guardrails. Fear dissolves through mastery.
Proof
80%+ content completion. 8% closed-won lift.
Seagate Systems Selling Academy, global sales team. Built by drafting the skeptics, designing for the workflow, and measuring execution.