# Move AI Adoption from Resistance to Results | Gleb Tsipursky

> AI adoption stalls when leaders treat it like a normal software rollout. Gleb Tsipursky shows how to clear the emotional blockers before you scale.

Guest: Gleb Tsipursky  
Host: Jim Rembach  
Published: 2026-11-18  
Video: https://www.youtube.com/watch?v=9SBaZvfPgng  
Source: https://www.fastleader.net/move-ai-adoption-from-resistance-to-results-gleb-tsipursky

## What you'll learn

- Why a CRM-style rollout produces pilot purgatory instead of P&L impact
- How confirmation bias, anchoring, and executive fear of messing up stall the leadership team
- The five middle-manager profiles that make or break adoption
- The eight employee attitudes, and why the AI alarmist is often the largest group
- The first move for each of the three emotional blockers

## AI Adoption That Reaches the P&L: Why Psychology Beats a Software Rollout with Gleb Tsipursky

Companies are buying AI like crazy and adopting it like molasses. The number named in this episode is blunt: 95% of enterprise pilots show no measurable P&L impact. The tools are in the building. The results are not.

**AI adoption** fails here because leaders treat it like the last software project. A CRM or an ERP gets a checklist, a training class, and an SOP. AI does not work that way. People have to talk to it, test it on their own work, and stay willing to look clumsy while they learn.

Gleb Tsipursky is the author of *The Psychology of AI Adoption at Work: From Resistance to Results*, published by Georgetown University Press. The profiles in this episode come from more than 100 consulting projects, more than 600 focus-group responders, and more than 10,000 survey forms, plus his Harvard Business Review article on middle managers. This conversation is about what to do before you scale the pilot.

## Executive Summary: Key Insights

- **A normal rollout misses the real bottleneck.** Training and SOPs do not create the buy-in AI needs, because each person has to adapt the tool to a real workflow.

- **Leaders stall on old biases.** Confirmation bias and anchoring keep the last technology playbook in place, while Gleb cites technology anxiety in 94% of technology leaders.

- **Three emotional blockers decide the scale-up.** Fear of job loss, threat to professional identity, and shame or social stigma have to be faced before a pilot can spread.

## Why Treating AI Like a Normal Technology Rollout Fails

Gleb's diagnosis is direct. Companies are treating AI as a normal technology. They run the same change plan they would use for a finance system or an HR system: announce it, train it, hand out the checklist, and wait for use.

That plan assumes the tool is standard. AI is not. He calls the work natural language programming. Each employee has to talk to the tool, figure out where it helps, and change the workflow. Leaders who skip that step also skip motivation. They are not focusing on people's psychology, and they are not focusing on getting people engaged.

He names three big emotional blockers and says they have to be addressed before AI can be adopted well. Training that ignores them teaches a skill people will not use. The pilot stays a pilot.

The same pattern shows up as a [Drift Tax](https://executionarchitects.com/glossary/drift-tax/): the company pays for the tool, the licenses, and the launch, then pays again when the work never changes. That repurchase of performance is the subject of [The Drift Tax: Why Organizations Pay Twice for the Same Performance, and How to Stop](https://thedrifttax.com/), which cites this body of research.

## The Biases That Keep Leaders on the Old Playbook

Gleb ties a 94% technology-anxiety finding to two biases. The first is confirmation bias. Leaders expected wide acceptance, so they look for proof that the old rollout should have worked. The second is anchoring. They are anchored to how the last system was adopted, so they copy that pattern onto AI.

He also names the fear under the executive team. He jokes about FOFU, the fear of messing up, because a bad AI bet can hit a bank account, a reputation, and a job. Culture slogans do not touch that fear. He says biological safety has to come before psychological safety. Those threat signals are [FONE Factors](https://executionarchitects.com/glossary/fone-factors/): status, certainty, and fairness are already in the room when someone hears that AI might replace them.

Rank-and-file resistance, he says, is a different set of blockers. Leaders asked him about leadership bias. He answered that first, then came back to the emotions on the floor.

## Five Middle-Manager Profiles That Gatekeep Adoption

Gleb calls middle managers the gatekeepers. They can make or break AI adoption. From most resistant to most enthusiastic, the five profiles are:

- **AI skeptic.** Worried about job loss, data privacy, output quality, and the human side of the change. Active doubt blocks the rollout.

- **Wait-and-see traditionalist.** Less alarmed, more passive. They want more proof, more memos, and a tool that works by magic. Inaction is the block.

- **Cautious implementer.** Open to AI, and serious about use cases, guidelines, escalation, and measurable outcomes. Strong at scaling a practice. Weak at creating the first enthusiasm.

- **Enthusiastic experimenter.** Tries tools, shares prompts, and runs informal demos so AI feels practical.

- **AI catalyst.** Future-oriented and inspiring, and able to push past the organization's comfort and past governance if nobody slows them down.

The research basis he gives is chapter 5 of the book and the Harvard Business Review article he said was published on September 1. Executives do not need every manager to become a catalyst. They need to know which profile is sitting on the pilot, because a building full of wait-and-see traditionalists will look like agreement and produce no use.

## Eight Employee Attitudes, Starting with the AI Alarmist

Employee attitudes are finer-grained than the manager set. Gleb starts at the resistant end. The **AI alarmist** is often the largest group. The fear is job loss first, then blame for errors. These are people who refuse the tools.

He walks the rest of the curve in order:

- **Pragmatic resistor.** Also refuses the tool, but the threat is professional identity and existing process, not only the job.

- **Skeptical observer.** Wait-and-see with doubt. They ignore a rollout that is not mandated.

- **Apathetic bystander.** Close to the observer, and not moved to explore.

- **Reluctant adopter.** Complies when told. Curiosity and innovation cannot be mandated, so use stays thin.

- **Cautious optimist.** Sees the upside and holds the guardrails. Useful, and not the source of new workflows.

- **Efficiency seeker.** Uses AI to streamline real tasks and will experiment to make the job faster.

- **AI evangelist.** The employee parallel to the AI catalyst. High energy, latest tools, and pressure on management to roll AI out broadly.

The split he cites from Pew is the size of the problem: 52% of people are more anxious than excited about AI, and only 9% are more excited than anxious. A launch aimed at the excited 9% will not move the anxious half.

## Three Emotional Blockers, and the First Move for Each

### Fear of job loss

Gleb starts with an executive choice. To the extent AI creates efficiency, aim that efficiency at growth rather than at cutting jobs. He cites Stanford research that more AI-forward companies in the same industry grow the top line over 9% faster and grow headcount over 6% faster. They do the same work with fewer people per unit, then take share, and still need people to take that share. Some of his clients commit to being job neutral: they will not fire people because an AI tool made a task faster. Jobs will change. The commitment is about what the company will not do.

The second move belongs to middle managers. People believe that using the tool trains their replacement. The conversation he wants is the opposite, and it has to stay honest. People who learn the tools become the productive ones the company keeps. He cites a PwC finding that domain experts with AI skills command a 69% wage premium. Outside the company, the skill is worth more too.

### Professional identity

The second blocker is meaning. A CRM lookup does not threaten someone's craft. AI can. Gleb's move is to find the work people love and the frustrations that block that work, then build prompts and agents for those jobs. The tool takes the friction and leaves the identity. Training is more likely to stick because it does not feel like an attack on the person.

### Shame and social stigma

The third blocker is colleagues who punish AI use. Alarmists and pragmatic resistors set a social cost. Gleb's move is modeling. Executives and middle managers use the tools in public on the work they already own: P&L analysis, board presentations, major sales meetings, email, and performance reviews. Norms move when the people with status go first.

Until those three are handled, scaling the pilot spreads the failure. He puts it plainly: you can lead a horse to water, but you cannot make it drink. If you do not address the emotional blockers first, you will be lost when you try to scale the AI pilot across the enterprise.

## Upgrade the Operating System Before You Scale

The close of the conversation is a systems point, not a motivation speech. Organizations try to implement AI while they are still conducting business in a [pre-AI operating system](https://executionarchitects.com/glossary/pre-ai-operating-system/). The old system asks people to remember, to comply, and to push through resistance with willpower. That will not carry a tool that requires experiment, talk, and local adaptation.

Getting from resistance to results is architecture. A [Leadership Execution System](https://executionarchitects.com/glossary/leadership-execution-system/) is the set of prompts, roles, and feedback that makes the new behavior the normal one: the growth commitment, the manager conversation, the identity-safe use cases, and leaders using the tool where everyone can see it. Fast Leader Show is the official podcast of Execution Architects. Execution is an outcome of system architecture, not human willpower.

Asked which single skill he would take back to age 25, Gleb does not pick a prompt library. He picks emotional self-management. Emotions motivate the leader and everyone else. A leader who cannot see that will keep launching pilots the workforce is too afraid to use.

## Key Tips for AI Adoption That Survives the Pilot

- **Do not copy the CRM plan.** If the rollout is a checklist and a class, it is the wrong plan for a tool people have to talk to.

- **Name the profile before you coach the person.** A wait-and-see traditionalist and an AI catalyst need different moves.

- **Aim the first tools at loved work and daily friction.** Identity stays intact when the tool removes what people already hate.

- **Model use on real management work.** A slide about adoption does less than a leader who drafts the board deck with the tool and says so.

- **Hold the scale-up.** Enterprise rollout before the three blockers are faced repeats the 95% pattern.

## Pros and Cons of a Psychology-First AI Rollout

**Pros**

- Matches how the tool is actually learned: talk, test, adapt

- Gives middle managers a role other than passive approval

- Puts a growth commitment next to the job-loss fear instead of a slogan

- Makes later training more likely to stick

**Cons**

- Slower than a mandate, and the excited minority will want to sprint

- Requires executives to make a real statement on jobs, not a vague culture line

- AI catalysts can still outrun governance if nobody pairs them with cautious implementers

- Individual manager conversations take time the old playbook never budgeted

## Top 10 Proven Steps to Move AI Adoption from Resistance to Results

- **State the P&L gap out loud.** Start from the pilots that show no measurable impact, not from the tool list.

- **Retire the normal-technology plan.** Drop the assumption that a checklist and an SOP will create use.

- **Audit leadership bias.** Look for confirmation bias and anchoring to the last system rollout.

- **Name executive fear.** Reputation, role, and results are in play. Pretending otherwise blocks psychological safety.

- **Sort the middle managers.** Skeptic, traditionalist, cautious implementer, experimenter, catalyst. Coach the profile you actually have.

- **Sort employee attitudes.** Expect the AI alarmist group to be large. Do not design the launch for the 9% who are already excited.

- **Choose growth over job cuts.** Make the executive commitment before you ask anyone to learn the tool. Job neutral is a real option he uses with clients.

- **Hold the one-on-one.** Managers tell people the truth: the skill raises their value inside and outside the company. Silence lets the replacement story win.

- **Protect the craft.** Point the first prompts and agents at frustrations and at the work people want to keep.

- **Model, then scale.** Leaders use AI on their own P&L, decks, sales meetings, and email. Only then do you spread the pilot.

## Memorable quotes

> The problem is that companies are treating AI as a normal technology.

> They're not really focusing on the buy-in. They're not focusing on people's psychology.

> You need to get people to be much more motivated, engaged, enthusiastic about AI tools.

> They can make or break AI adoption.

> By using AI tools, you're not training your own replacement.

> If you don't address the emotional blockers first, you will be lost when you try to scale up the AI pilot across the enterprise.

> Never trust your gut.

## FAQ

**Why do most enterprise AI pilots fail to show P&L impact?**

Gleb says leaders treat AI like a normal technology project. The tool is easier to click than a legacy system and harder to adopt, because each person has to experiment. Pilots stall when emotional blockers are ignored. He points to MIT research that 95% of AI pilots do not show the impact leaders expected.

**What are the three emotional blockers to AI adoption?**

Fear of job loss, threat to professional identity, and shame or social stigma from colleagues who punish AI use. Each one needs a different first move: a growth commitment, tools aimed at loved work, and leaders modeling their own use.

**How should executives get middle managers to support AI?**

Identify which of the five profiles you have. Skeptics and wait-and-see traditionalists block by doubt or by inaction. Cautious implementers scale a practice once enthusiasm exists. Experimenters and catalysts create energy, and catalysts need governance so they do not outrun the rest of the company.

**Should a company promise not to cut jobs because of AI?**

Gleb recommends an executive commitment to growth rather than to cutting jobs with the efficiency AI creates. Some of his clients commit to being job neutral. He is clear that jobs will change. The commitment is that people are not fired because a tool made the old task faster. He cites Stanford figures that more AI-forward firms grow revenue and headcount faster than peers.

## Episode timestamps

- 0:00 Why AI spending is not showing up in the P&L
- 1:54 Why a normal technology rollout fails for AI
- 8:10 The biases stalling leadership on AI
- 16:48 Five middle-manager profiles that gatekeep adoption
- 25:58 Eight employee attitudes toward AI
- 39:37 First step: face the fear of job loss
- 44:21 Second blocker: professional identity
- 46:44 Third blocker: shame and social stigma
- 49:40 Hump Day Hoedown
- 53:07 The one skill Gleb would take back to age 25

## Research and citations

- [Gleb Tsipursky, The Psychology of AI Adoption at Work: From Resistance to Results, Georgetown University Press. Free sample he named on the episode: disasteravoidanceexperts.com/ai-book.](https://disasteravoidanceexperts.com/ai-book)
- Harvard Business Review article on middle-manager psychographic profiles, which he said was published on September 1. Chapter 5 of the book covers the same profiles.
- Research base he stated: more than 100 consulting projects, more than 600 focus-group responders, and more than 10,000 survey forms.
- MIT research, as cited in the episode: 95% of AI pilots fail to deliver the P&L impact leaders expect.
- Pew, as cited: 52% more anxious than excited about AI, 9% more excited than anxious.
- Stanford, as cited: more AI-forward companies grow top line over 9% faster and headcount over 6% faster than less AI-forward peers in the same industry.
- PwC, as cited: domain experts with AI skills command a 69% wage premium.
- [The Drift Tax: Why Organizations Pay Twice for the Same Performance, and How to Stop](https://thedrifttax.com/)
- [Pre-AI Operating System](https://executionarchitects.com/glossary/pre-ai-operating-system/)
- [Leadership Execution System](https://executionarchitects.com/glossary/leadership-execution-system/)
- [FONE Factors](https://executionarchitects.com/glossary/fone-factors/)
- [Drift Tax](https://executionarchitects.com/glossary/drift-tax/)
