The senior leadership team that is not engaging with AI is rarely dismissing it. They are deferring it; and they are deferring it because the conversation has not yet reached the level of specificity that requires a decision.
“We should probably do something about AI” is a comfortable level of agreement that produces no action.
“We are going to spend $X on a Phase 1 AI Foundations engagement starting in [month]; and [person] will own it; do we agree?” is a decision.
The difference between these two statements is the difference between a senior team that acknowledges AI is important and one that has committed to doing something about it.
This article is for the person in the room who is already convinced and needs to move the rest of the leadership team from acknowledging AI to committing to it.
The approach is specific: understand the resistance type; make the case in the language each resistance type responds to; and close to a specific decision rather than a general agreement.
The four resistance types: what each one actually needs to move
Resistance Type 1: The Skeptic (most common in companies with established processes)
Who they are: the partner or VP who has seen technology initiatives come and go without producing the promised results. They have a pattern-matching response to “AI will change everything” that is grounded in experience. They are not wrong to be cautious; they are applying the right prior to the wrong situation.
What they say:
“We tried a workflow automation initiative three years ago and it didn’t stick. This feels like the same thing.”
What they actually need: specific evidence that this situation is different from the past technology initiatives that did not stick.
Not general claims about AI’s capability; specific answers to why the previous initiatives failed and how this approach addresses those failures.
The most effective response:
- Acknowledge their experience directly: “You’re right that the previous initiative did not stick. Here is specifically why: there was no context layer; no documented workflows; and the engagement partner left after delivery with nothing running.”
- Name the specific differences: “The approach we are considering builds the foundation first; documents every workflow; installs a feedback loop; and does not exit until the system owner can maintain it independently.”
- End with evidence: “Here is the output the previous approach produced and here is the output the same prompt produces with the context pack loaded.”
The Skeptic responds to specificity; not enthusiasm.
Resistance Type 2: The Risk Avoider
Who they are: the senior leader who is primarily worried about downside; the team anxiety; the client perception; the distraction from core operations; the cost of a failed initiative.
They are not against AI. They are against risk that cannot be managed.
What they say: “I’m worried about how the team will react.” / “What if clients find out we’re using AI?” / “We can’t afford a distraction right now.”
What they actually need: a specific plan for managing each risk they are naming. Not dismissal of the risk; an answer to it.
The most effective responses by risk type:
| Risk raised | Specific answer |
|---|---|
| Team anxiety | Describe the involvement-first implementation approach. Name the communication plan. Show the specific job security language that will be used. |
| Client perception | Share the disclosure framework and the specific language. Show that proactive disclosure is a trust-building tool; not a risk. |
| Distraction from operations | Name the founder’s time commitment (4–6 hours over two weeks for Phase 1). Show that the operational distraction is bounded and specific. |
| Cost of a failed initiative | Describe the phase gates that prevent investment in Phase 3 before Phase 2 is proven. Show that the downside is limited to Phase 1 if Phase 1 does not work. |
The Risk Avoider responds to managed risk; not dismissed risk.
Resistance Type 3: The Deferrer (the most common non-resistance resistance)
Who they are: the senior leader who agrees AI is important but consistently supports deferring the decision; after the next client cycle; after the next financial review; after the team stabilises. They are not opposed; they are perpetually “not the right time.”
What they say: “I agree we should do something. Can we revisit this next quarter?”
What they actually need: a specific cost-of-delay argument. Not abstract urgency; a concrete description of what the company is giving up each quarter it defers.
The most effective response:
Make the deferral cost concrete:
“If we defer to Q3; we are twelve weeks further behind a competitor who started in Q1. Our Phase 1 will not be complete until Q4. Our team will not be trained until Q1 next year. The first workflows will not be running until Q2 next year; eighteen months after a competitor who started this quarter.”
Then offer the constrained choice: “The decision today is not whether to build AI capability; it is whether to start this quarter or next. What would we need to be true to choose this quarter?”
The Deferrer responds to specific delay costs; not general urgency language.
Resistance Type 4: The Proprietor (most common in professional services)
Who they are: the senior partner or founder who believes their professional expertise is the primary competitive advantage and perceives AI as a threat to that identity.
“What we sell is judgment”; and the implicit concern is that AI undermines the value of that judgment.
What they say: “Our clients pay for our thinking; not for AI outputs.” / “The quality of what we produce comes from experience; not software.”
What they actually need: a reframe that positions AI as an amplifier of their expertise; not a replacement for it.
The most effective response:
“We agree. The judgment is the product. AI handles the desk work; the research; the first drafts; the data assembly; so your time goes entirely to the judgment layer. The client gets more of what they are paying for; not less. You produce more of the thinking and less of the supporting work.”
This argument works because it is accurate. The Proprietor’s concern is valid; and the response addresses it honestly.
The show-before-you-tell strategy: the most effective single intervention
Why showing works better than telling
The leadership team that has never seen AI produce a company-specific output is evaluating AI based on what they have read and heard; which is almost entirely enterprise-scale examples that do not feel relevant to a $15M company.
The ten-minute demonstration of AI producing a proposal draft; a client briefing; or an operational report in the company’s voice; for the company’s client type; eliminates three objections that a presentation slide never could.
How to structure the demonstration
Step 1: Choose the right output type
The output type should be something the most senior person in the room recognises as valuable immediately.
| Leadership profile | Best demonstration output |
|---|---|
| Professional services (partners; account leads) | Proposal draft or client recommendation letter |
| Operations-focused (COO; ops director) | Operations summary or financial briefing |
| Sales-focused leadership | Prospect research brief or follow-up email sequence |
Step 2: Prepare with a full context pack loaded
Do not demonstrate AI without context loaded. The demonstration without context shows what the team already assumes. The demonstration with the company’s context shows what they need to see.
Step 3: Run it live in the meeting; not as a pre-recorded example
Live demonstrations are more credible than pre-recorded ones because the audience knows the output was not cherry-picked.
Take the specific input from the room: “use this client” or “use this project”; and run it live.
Step 4: Ask two questions after the output appears
- “How long would this have taken manually?” (Makes the time value visible)
- “What would you change about this draft?” (Invites professional evaluation; not technology assessment)
The second question is the most important. The leader who is editing the output is no longer evaluating AI as a technology; they are using it. That shift in posture produces a different conversation.
The decision question: how to close to commitment rather than agreement
The gap between agreement and commitment
Most AI discussions in senior leadership meetings end in agreement; “yes; we should do something about AI.”
This agreement is comfortable and produces no action. The conversation needs to end in a specific commitment; not to AI in general; but to a specific investment; a specific timeline; and a specific owner.
The structure of a decision question
A decision question has three elements:
- A binary or constrained choice (forces a specific decision rather than a general discussion)
- A specific stake (what changes depending on which choice is made)
- A “what would need to be true” component (invites engagement with the conditions rather than the conclusion)
The most effective AI decision question
“Do we invest in Phase 1 AI Foundations this quarter or next quarter? [pause] What would we need to be true to choose this quarter?”
Why this question works:
- Creates a specific binary choice rather than a general discussion
- Implies both choices are valid; reducing the pressure that produces defensive responses
- Opens with the conditions for the faster choice; which usually produces a list of answerable concerns rather than continued deferral
How to handle the most common response (“what would it cost?”)
Have the specific number ready: “Phase 1 is [specific cost] over [specific time]. It produces [specific deliverables]. The first measurable outcome appears within [specific timeframe].”
The specificity of the answer signals preparedness and reduces the uncertainty that drives cost-based deferral.
The pre-meeting private conversation
The single most effective preparation for a leadership AI conversation: a five-minute private conversation with the most senior person who has not yet formed an opinion.
Their neutrality in the group conversation creates permission for others to defer.
Their private engagement; “I wanted to share what I’ve been thinking before Tuesday’s meeting and get your read on it”; often produces a very different group conversation than the one that would have happened without it.
After the commitment: what the first thirty days should produce
The commitment made in the leadership meeting is the beginning of the implementation; not the end of the persuasion.
Commitments made without immediate; visible progress markers erode within two to four weeks as operational demands compete with the new initiative.
The thirty-day visibility plan
Week 1: The founder’s visible action
The first visible action should happen before the following Monday. Options:
- The first context pack writing session is scheduled (even if not yet complete)
- The AI system owner is named and announced to the team
- The first tool subscription is upgraded from personal to shared
Any of these signals that the commitment produced action; not just minutes.
Weeks 2–3: The first output the room sees
The leadership team sees their first AI-produced output. It should be from the company’s actual operations; a draft of a real document; a response to a real client situation; a report from real operational data.
The output demonstrates that the investment they committed to is already producing something.
Week 4: The first metric
A specific number is reported back to the leadership team: the acceptance rate on the first workflow; the time saved in the first two weeks; the number of workflows documented.
The metric does not need to be impressive. It needs to be real. A real number signals that the implementation is being measured; not just run.
Common questions on making the leadership AI case
”What if the most senior person in the room is the strongest skeptic?”
Use the Skeptic approach from above; but with additional preparation.
Before the meeting: identify the specific technology initiative they are pattern-matching to and prepare the specific “here is why this is different” argument with evidence for that specific comparison.
In the meeting: address their skepticism first; before the broader group discussion. “I know you’ve seen this kind of initiative not stick before. I want to address that directly.” Engaging the strongest skeptic first; respectfully and specifically; prevents their concern from becoming the group’s reason to defer.
”How do I build a business case when I don’t have internal ROI data yet?”
Use the operational specifics rather than ROI projections.
“A proposal that currently takes three days will take three hours with AI assistance. At the volume of proposals we produce; that is [specific hours] per month redirected to client relationship work.” This is not ROI; it is a specific operational claim the leadership team can evaluate and believe.
Abstract ROI projections (37% productivity improvement) are dismissed because they feel like vendor marketing. Specific operational claims (three days to three hours on proposals) are believed because they are verifiable.
”What if the leadership team agrees in the meeting but the momentum dies within two weeks?”
The thirty-day visibility plan above is the answer to this. The commitment that does not produce a visible action in week one has already lost momentum.
Specifically: the named AI system owner is the single most important visible action. A name on a role means accountability. The absence of a named owner means the commitment is still theoretical.
”What is the minimum viable AI commitment a leadership team can make that still produces movement?”
The minimum viable commitment: “We will invest in Phase 1 AI Foundations; [person] will own the initiative; [start date].”
This commitment names the investment; names the owner; and names the date. Without all three; the commitment is not a commitment; it is a general agreement to explore.
Want the leadership conversation structured: with the demonstration ready, the decision question prepared, and the Phase 1 plan specific enough to commit to in the meeting?
The leadership team that does not take AI seriously is usually one conversation away from taking it seriously.
That conversation works when it ends in a specific decision rather than general agreement; uses the evidence type each resistance profile responds to; and includes a ten-minute demonstration of AI producing something the room recognises as useful.
Agreement without visible first actions; an early output the team can see; and a first metric within thirty days becomes another initiative that peaked in the meeting and quietly disappeared by the following month.
Path one: run the demonstration before the next leadership meeting. Load the company’s context pack into a Claude Project. Choose the output type most likely to resonate with the most senior person in the room. Run it live in the meeting on a real current task. The output does more work in ten minutes than a deck does in forty.
Path two: bring in a partner. Phos AI Labs has had this leadership conversation dozens of times; with the demonstration prepared; the Phase 1 plan specific; and the decision question structured. The first conversation produces the specific plan the leadership team needs to move from agreement to commitment. We have run 400+ AI engagements. Clients include Zapier, Coca-Cola, Medtronic, Dataiku, and American Express. Thirty minutes, no deck. Start here.