Leaders ADAPT teaching package

Teaching AI readiness: a free session plan for managers and leadership teams

A ready to run class or workshop that teaches managers to decide where AI belongs in their work, and where it does not, before anyone buys another tool. Free to use under CC BY 4.0. Participants walk out knowing which wave of AI adoption they are actually in, with five real inefficiencies triaged into AI, process or people, and one tiny sliver picked to pilot.

20, 45 or 90 minutesFree test includedPrintable worksheetCC BY 4.0Vendor neutral

Who this is for

Instructors of MBA courses in digital transformation, technology management and operations, university courses on the future of work, corporate L&D teams running AI upskilling for managers, and team leaders who want one practical session with their own team before the next AI initiative starts. Participants do not need to be technical. The method is about judgment, not tools, and works the same whichever AI products your organization allows.

What participants leave with

A readiness level and scores on Placement, Restraint, Ownership and Momentum from the free test, and the wave of AI adoption they are actually in, which is often not the one their organization talks about.
Five real inefficiencies from their own work, each triaged as an AI, process or people fix with a reason, after the intern rule has been applied.
One tiny sliver to pilot in the next two weeks, with a named human reviewer and a rule for when the prompt becomes a reusable skill.

Pre-work (optional, 10 minutes)

Ask each participant to keep a short list for one week: every time something at work felt slow, repetitive or annoying, write one line. Not solutions, just the irritation and roughly how often it happens. They bring the list. If nobody does it, give 5 minutes at the start of the session to list the last two weeks from memory and their calendar.

Ready, fire, aim: AI adoption as a judgment problem

Andreas Pettersson ran an AI company as CEO from 2016, before generative AI went mainstream, and now teaches leadership teams to adopt it. His position: AI adoption is a mindset shift before it is a tooling shift. Most organizations aim, aim, aim, with working groups and strategy decks and almost no usage. The seven ideas below are how he moves a team from talking about AI to using it well, without bolting a tool onto a problem that was never a tooling problem.

The three waves of AI adoption: Wave one is personal productivity: individuals get faster at their own work. Wave two is cross department efficiency: the gains move between teams and into process. Wave three is AI inside the product or service you sell. Most companies want to start at wave three, because that is what sounds like valuation. That is the wrong order. This is an efficiency wave first, a margin play: faster execution at a lower cost. The rule: find out which wave a team is actually in, not which one it is talking about. Almost nobody has finished wave one.
Ready, fire, aim: Aim, aim, aim is endless discussion and no usage. The cure is to fire: use the tools heavily, be scrappy, iterate fast. Then aim again: review what came out. There are two failure modes and a leader has to hold both. Perfectionism kills adoption. Unreviewed output shipped to customers kills quality.
Discovery before prescription: Before a leader decides anything about AI for a team, Andreas runs a short one to one round with the people doing the work. Every person gets the same question: where are the inefficiencies, and is the fix AI, a process, or a cheaper pair of hands? No solutions in the call: this is discovery, and the prescription comes later. At the end he asks each person not to compare notes with colleagues until the round is finished, so the answers stay independent. Then act on patterns, not on single voices. If most people name the same problem, the answer is clear. If everyone names something different, you do not have a diagnosis yet.
AI, process or people, and the intern rule: For every inefficiency, ask three questions before you buy anything. Is the fix a process: a clearer handoff, a template, a meeting that changes or disappears? Is it people: delegation, a cheaper pair of hands, a role nobody owns? Or is it genuinely AI? Do not run the triage in favour of AI. Putting AI on top of a broken process or an unclear owner treats the symptom, not the root cause. Ask what the goal of the work is, not what the task is, until the real problem shows up. The guardrail on the whole triage is the intern rule: organizations often automate something, or put a tool in place, when it would have been cheaper to have an intern click the buttons. For mundane, non time critical work, ask whether a junior person with a checklist would solve it faster and cheaper. If yes, that is the answer.
Isolate a tiny sliver: You cannot do Big Bang. An organization that is not ready should not try to roll AI across every system and archive at once. Pick one narrow, repeated piece of work, prove value there, then expand. This is matching the scope to the maturity of the team, not a limit of the technology.
Human in the loop: AI is a force multiplier, not a final author. Every output that reaches a customer must meet the company's own standard for accuracy, voice and thinking. The working rule: assume it is always wrong, so verify everything before you ship it. Every sliver needs a named person who reviews the output before it leaves the building.
Prompts versus skills: A prompt is a one time ask. A skill is a reusable teammate that improves. Think of a prompt as an intern you brief once and never see again, and a skill as an intern you build up over time. The promotion rule: the moment you copy and paste the same prompt a second time, turn it into a saved, shared, reusable instruction. That is how individual tricks become team knowledge, and how a team moves from wave one toward wave two.

The phrase Ready, Fire, Aim is the title of a book by Michael Masterson; Andreas Pettersson uses it as a pacing rule for AI adoption. The three waves, discovery before prescription, the AI, process or people triage, the intern rule, isolate a tiny sliver, the human in the loop rule and the prompts versus skills promotion rule are his.

Want the system underneath? The 5-Minute Leader

The 5-Minute Leader is not an AI course. It installs the leadership rhythm AI adoption runs on: clear handoffs, honest 1:1s, accountability and a fixed weekly cadence. A pilot with no owner, no review slot and no weekly check stalls whatever the tool. This package teaches where AI belongs. The book teaches the operating rhythm that keeps a team using it, step by step.

The Delegation Protocol: handing over outcomes instead of tasks, SCQA framing and three explicit authority levels, the same discipline a good AI brief and a named reviewer need
The 1:1 Protocol, where you find out what people are actually struggling with, which is where real adoption shows up or quietly stalls
The Accountability Protocol, for the moment a committed pilot slips and someone has to have the conversation
The Rhythm Protocol: a daily stand-up and a weekly team meeting agenda, a fixed place to review every pilot every week

Get the full 5-Minute Leader system

Session plans

Pick the length that fits your slot. Every version uses the same test, framework and worksheet.

20 minute version

Minutes	What happens
0 to 3	Frame: ask for a show of hands. Who has sat in a meeting about AI in the last month? Who used AI on real work in the last 24 hours? The gap between the two counts is the session.
3 to 7	Everyone takes the free AI leadership readiness assessment on their own device. Note the level and the lowest of the four dimensions.
7 to 11	Teach the three waves and Ready, fire, aim. Each person ticks the wave map on the worksheet and writes the wave their organization talks about next to the wave they are actually in.
11 to 17	Teach AI, process or people and the intern rule. Each person triages three items from their pre-work list in the triage table. Pairs challenge one AI answer each: would an intern or a process fix it?
17 to 20	Each person circles one AI item as their tiny sliver, names the human reviewer, and says it out loud in one sentence.

45 minute version

Minutes	What happens
0 to 5	Frame with the two show of hands questions from the 20 minute version. Write both numbers on the board.
5 to 10	Everyone takes the free AI leadership readiness assessment. Show of hands per lowest dimension.
10 to 17	Teach the three waves and Ready, fire, aim. Everyone completes the wave map on the worksheet. Ask three people to read out the gap between the wave they talk about and the wave they are in.
17 to 25	Read the fictional case aloud. Small groups answer discussion questions 1 to 3.
25 to 37	Teach AI, process or people, the intern rule and isolate a tiny sliver. Each person fills in the triage table with five items from their pre-work list. Pairs swap sheets and challenge every AI answer with two questions: what is the goal of this work, and would an intern with a checklist do it cheaper?
37 to 42	Teach human in the loop and prompts versus skills. Each person completes the tiny sliver box: what, who reviews, what good looks like in two weeks, and when the prompt gets promoted to a skill.
42 to 45	Each person says their sliver and their reviewer out loud. Close with discussion question 7.

90 minute version

Minutes	What happens
0 to 8	Frame with the two show of hands questions. Then ask: what has your organization already decided about AI, and how much of that decision has produced daily use? Collect answers without judging them.
8 to 15	Everyone takes the free AI leadership readiness assessment. Pairs compare levels and lowest dimensions and discuss whether the result matches what their manager would say about them.
15 to 28	Teach the three waves and Ready, fire, aim. Wave map on the worksheet. Groups of four agree on the single piece of evidence that would prove a team had finished wave one.
28 to 43	Fire round. If participants have access to an approved AI tool, each person uses it for 8 minutes on one real, non confidential piece of their own work from the pre-work list: a draft, a summary, a checklist. No polishing. Then aim: 5 minutes to review the output against the human in the loop rule and mark every error, invented fact or wrong tone. If the room has no tool access, run it as a demo on screen with the instructor's own work and have the room do the review.
43 to 55	Read the fictional case. Groups work discussion questions 1 to 5. Each group presents its triage of Fernhill's five problems.
55 to 72	Triage table on the worksheet with five items each. Pairs challenge every AI answer with the goal question and the intern rule. Then groups of four pick the one sliver from their combined lists they would fund first and explain why it is the smallest useful one.
72 to 82	Teach prompts versus skills. Each person writes the prompt they used in the fire round as a reusable instruction: who it is for, the context it needs, the standard the output must meet, and who reviews it. Pairs test whether a colleague could run it without asking a question.
82 to 90	Tiny sliver box on the worksheet. Each person says their sliver, their reviewer and their two week check date out loud, plus the three colleagues they will ask the discovery question before the pilot starts. Close with discussion question 7.

The free test participants take

Participants take the free AI leadership readiness assessment on their own phone or laptop. It takes a few minutes. It asks for a name and email to send the full result, so tell participants that in advance. Results go only to the person who took it, so nobody has to share a score unless they choose to.

Link to give participants (paste it into your slides or learning platform):

https://www.leadersadapt.com/ai-readiness-assessment/?utm_source=teaching-package&utm_medium=referral&utm_campaign=teach-ai-readiness

The case: the AI working group that never shipped

Fictional case, written for teaching. Names, companies and numbers are invented.

Fernhill Engineering is a fictional 90 person civil engineering consultancy. After a board meeting in January, the managing director, Sofia Brandt, set up an AI working group of six senior staff. Its brief was to design an AI powered client portal that answers project questions around the clock. Sofia believes the portal will set Fernhill apart when the firm is valued.

Eight months later the group has met 14 times, compared four vendors and written a 30 page requirements document. Nothing has launched. Outside the group, a handful of junior engineers use free AI tools on their own to draft reports. Nobody knows which tools, and nobody reviews what comes out. Last month one of those drafts reached a client with a load figure that did not exist in the source calculation. A senior engineer caught it the next day.

The operations manager, Daniel Achebe, has his own list. Site visit notes are typed up twice, once for the report and once for the invoice. Every proposal is written from a blank page, although most of them reuse the same eight sections. The weekly project meeting runs 90 minutes because nobody owns the status tracker, so it gets rebuilt live in the meeting. Timesheet reminders go out by hand every Friday. New starters spend their first two weeks asking the same questions about templates and file locations.

Sofia has asked the working group for a final recommendation by the end of the quarter. The group is leaning toward a larger vendor and a six month rollout across all projects and the full document archive.

The handout

One sheet per participant. Fill in Part 1 and Part 2 in the session, Part 3 before you leave. Download the Word version

Teaching packageTeaching AI readiness

Leaders ADAPT | CC BY 4.0leadersadapt.com/teach-ai-readiness/

Which wave are you in? Plus the AI, process or people triage

PART 1. WHICH WAVE ARE YOU REALLY IN? Tick every statement that is true today, not planned. WAVE 1. Personal productivity [ ] I used AI on real work in the last 24 hours [ ] I use it at least daily, not only when I remember [ ] I have at least one prompt I reuse instead of rewriting [ ] Most people on my team use it daily on their own work [ ] Someone reviews AI output before it reaches a customer WAVE 2. Cross department efficiency [ ] At least one handoff between teams now runs with AI in it [ ] We have shared, reusable instructions (skills) that more than one person runs [ ] A process was redesigned, not just sped up [ ] We can name a cost or hours saved and who measured it WAVE 3. AI in what we sell [ ] Customers use something with AI inside it that we built or configured [ ] It has a named owner and a quality check The wave my organization talks about: 1 2 3 The wave we are actually in: 1 2 3 (You are in a wave only when every box of the wave below it is ticked.) The one missing box that matters most: ______________________________________________ PART 2. TRIAGE TABLE (five items from your list) For each one: what is the goal of the work, not the task? Then AI, process or people. Intern rule: would a junior person with a checklist do this cheaper? If yes, it is people. # Inefficiency (how often) | Goal of the work | AI / Process / People | Why 1 ________________________ | ____________________ | ______ | ____________ 2 ________________________ | ____________________ | ______ | ____________ 3 ________________________ | ____________________ | ______ | ____________ 4 ________________________ | ____________________ | ______ | ____________ 5 ________________________ | ____________________ | ______ | ____________ Partner check: which AI answer did your partner challenge, and did it survive? ______________________________________________ PART 3. ONE TINY SLIVER The one narrow, repeated piece of work I will pilot: ______________________________________________ Why this is small enough (one person, one task, under two weeks): ______________________________________________ Human in the loop. Who reviews every output before it leaves: ______________________________________________ What good looks like at the two week check: ______________________________________________ Check date: __________ Discovery first. Three colleagues I will ask the same question, one to one, before I start: 1. ________________ 2. ________________ 3. ________________ Promotion rule. The day I paste the same prompt a second time, I save it as a reusable skill. The skill will be called: ______________________

Teaching AI readiness: a session plan. Andreas Pettersson, Leaders ADAPT. Free to use under CC BY 4.0. https://www.leadersadapt.com/teach-ai-readiness/ The full system: 5minleader.leadersadapt.com/5ml

Discussion questions

Which wave is Fernhill actually in, and which wave is the working group designing for? What evidence in the case tells you?
Is the working group aiming, firing, or neither? What would firing look like for Fernhill in the next two weeks?
Triage Daniel's five problems into AI, process or people. Which ones fail the intern rule? Which one is not an AI problem at all, even though AI could make it faster?
The junior engineers are the only people at Fernhill actually firing. Should Sofia stop them, or learn from them? What does the human in the loop rule require either way?
Pick the tiny sliver you would start with at Fernhill. Why that one, and who reviews the output?
Which of your own repeated prompts should already be a skill? What stops people on a team from sharing what they have built?
What is the one decision about AI in your own work that you have been aiming at for too long?

Facilitator notes

The assessment asks for an email before it shows the full scorecard. Tell participants before they start. The session works without the score: the wave map and the triage table carry the learning on their own.
Data rule for any live AI use in the room: nothing confidential, nothing about customers or colleagues, only the participant's own non sensitive work and only in tools their organization allows. Say this before the fire round, not after. The same care applies to anyone who runs the discovery round back at work: explain the purpose up front, keep it to people they work with directly, and report patterns only, never who said what.
Expect the room to triage almost everything as AI on the first pass. That is the point. The debrief question that turns it around is the goal question: what is this work for? Ask it until the real problem shows up, usually an unclear owner or a missing process.
Skeptics will argue that AI output is unreliable. Agree, then hand them the human in the loop rule: assume it is always wrong and verify before you ship. Unreliable output is a review problem, not a reason to keep aiming.
If nobody volunteers in the fire round, run your own real task live on screen and let the room find the errors in the output. Finding the mistakes is the lesson, so a flawed result works better than a perfect one.
Keep the session vendor neutral. If the conversation turns into which tool is best, park it on a flip chart. Tools change every few months. The triage and the review rule do not.

Use, adapt and cite

This package is licensed under Creative Commons Attribution 4.0. You may copy, print, adapt and share it in any course, workshop or internal training, including paid ones, as long as you credit the source with a link.

Citation

Pettersson, A. (2026). Teaching AI readiness: a session plan. Leaders ADAPT. https://www.leadersadapt.com/teach-ai-readiness/

Link for your course page or syllabus

<a href="https://www.leadersadapt.com/teach-ai-readiness/">Teaching AI readiness: a free session plan for managers and leadership teams</a> by Andreas Pettersson, Leaders ADAPT (CC BY 4.0)

Using the package does not imply any endorsement of Leaders ADAPT by your institution, and none is claimed. Teaching with it? Tell us at andreas@leadersadapt.com and we will list your course on the teaching packages page.

Keep going after the session

If your tiny sliver needs an owner, a reviewer and a weekly check that actually happens, that is a leadership rhythm problem. The 5-Minute Leader installs that rhythm one protocol at a time.