Role 1
Leadership
The execs who set strategy and allocate budget. Without them, training hits the ceiling.
Home / Services / AI Workforce Training
Module · AI Workforce Training
Get your team using it Monday.
If your team is not using AI on Monday, the strategy is not landing. Three tracks for three audiences: the execs who set direction, the few people who build, and everyone else. Built on Ethan Mollick's Leadership, Lab, and Crowd, and built to run without me.
Book 30 minutes with Edwin
A module that usually runs alongside or after a build. See where it fits.
- Duration4 to 8 weeksBy team size and track mix
- TracksThreeLeadership, Lab, Crowd
- Reinforcement30 / 60 / 90Days after the last session
- PriceIn writingWithin 48 hours of the call
01 · The framework
Leadership,Lab, and Crowd.
Ethan Mollick at Wharton named three internal roles every company adopting AI needs, in May 2025. I build the training on that. The typical program trains only the Crowd, hits a 20% adoption ceiling, and stalls. The fix is training all three in sequence.
The "secret cyborg" lives here too. People use AI privately and never share it, because nobody signaled that it is safe to admit.
One rule decides where AI lands first: aim it at the chores your people dread, and leave their craft alone. Relief is what creates belief. A leader cannot mandate enthusiasm, but a leader can arrange for relief. Start there, and make your most tenured people the champions, because the floor trusts them more than any memo.
Role 2
Lab
A small builder team running experiments. Without them, there are no internal examples to teach from.
Role 3
Crowd
The 80% using AI in daily work. Without them, only the power users benefit.
02 · The tracks
Three tracks,in sequence.
Track 1 · Leadership
The executive layer
For the execs setting AI direction. Three-Horizon portfolio framing, budget defense language, and eval oversight that does not turn them into a bottleneck.
2 × 90 minPlus async reading
Track 2 · Lab
The builder team
For your 5 to 10 builders. Prompt engineering for production, eval setup in Braintrust or an equivalent, multi-model routing and cost control, and agent literacy from single-shot to multi-agent.
4 × 2 hrsPlus hands-on builds
Track 3 · Crowd
Everyone else
For ops, sales, customer success, and engineering. Daily-use patterns, and prompting for the real use case instead of the demo.
2 × 60 minPlus async references
03 · The curriculum
Tied toyour workflows.
Generic AI training flatlines after the trainer leaves. This curriculum lives in the knowledge base your team and your AI both read from.
Typical AI training
- Prompt workshops disconnected from real workflows.
- The trainer leaves and the team forgets.
- Vendor tutorials that lock you in.
This training →
- Built on your actual use cases. The exercises are your work.
- References in the shared knowledge base. New hires find them without asking.
- Adoption tracked at 30, 60, and 90 days. Measured against a baseline.
04 · The engagement
4 to 8 weeks.Five phases.
Me, an exec sponsor, and one Lab lead. Live workshops, recordings, and async material, with reinforcement at 30, 60, and 90 days. The length depends on team size and how many tracks you run.
Adoption baseline · Week 1
I interview 6 to 10 people across all three roles, find the secret cyborgs, and pick the Crowd workflows with the most hours in them.
Leadership track · Weeks 2 to 3
Two exec sessions. Lock the direction, the allocation, and the metrics, and signal that it is safe to admit AI use.
Lab track · Weeks 3 to 6
Four sessions with your builders on real internal use cases. The eval framework gets stood up here.
Crowd track · Weeks 6 to 8
Two sessions per team on use-case prompting, plus a path for contributing back to the knowledge base.
Reinforcement · 30, 60, 90 days
An adoption review at each checkpoint, with course corrections on what did not stick.
05 · Where it fits
A module.Beside the path.
My main path is three steps: the plan, the build, and keep me on. Training sits beside it and usually runs alongside or after a build. Once a workflow ships, the people who run it need to read the evals, catch the failures, and spot the next one. If you are not sure training is the gap, the plan will tell you. Sometimes the real gap is that nobody owns the program.
-
01
The plan: AI Program Diagnostic
Two weeks. The workflow map, the 90-day sequence, and the kill list.
Where the main path starts. Credited in full toward the build or a retainer within 30 days.
$2,5002 weeks -
02
The build: Scoped AI Build
The first lines of the plan, shipped into the stack your team already runs.
From $7,500 a workflow, two-workflow minimum.
From $15kPer build -
+
AI Workforce Training
This page. Leadership, Lab, and Crowd, so your team runs what shipped.
Usually alongside or after the build.
4 to 8 wksPriced in writing -
03
Keep me on, monthly
AI System Continuity keeps it running. The Fractional retainer owns the program and the next builds.
Fractional is 1 to 3 days a week, 6-month minimum.
From $12kA month, Fractional
06 · The outcome
The ceilingmoves.
Last session
- An adoption baseline, measured.
- A Lab that can run the next round of training.
- Crowd references in the knowledge base.
- Leadership with the checkpoints to flag a regression.
Six months →
- Adoption past the 20% ceiling and still climbing.
- Fewer secret cyborgs. The good work surfaces.
- The Lab ships internal tools without outside help.
- Wins measured in hours reclaimed a week, never in how much AI got used.
07 · Also included
Beyond theworkshops.
Included 01
Every session recorded
Yours to keep for new-hire onboarding, refreshers, and anyone who missed it live. Training scales without bringing me back.
Included 02
The playbook
Leadership, Lab, and Crowd documented for internal reuse, with the checkpoints, exercises, and talking points for the exec sponsor. Your Lab runs the next cycle.
Included 03
Four weeks of office hours
A weekly drop-in after the last session for live questions, prompt review, and workflow troubleshooting. It catches regressions before they set in.
08 · The guarantee
Adoption movesor I fix it.
Adoption gets re-measured at day 60. If the metric has not moved, the course correction is on me.
Day 60 re-measure
No movement, no charge to fix it
If the day 60 re-measure shows no movement from baseline on the agreed metric, I run the adoption baseline again (one session) and write a course-correction memo at no extra cost.
Exit
14-day mutual exit
Either party can end the training inside the first 14 days. Days worked are billed at the agreed day rate, with no further commitment.
09 · FAQ
Questions.
What is AI workforce training based on Leadership-Lab-Crowd?
A 4 to 8 week program that trains the three internal roles every AI-adopting company needs: Leadership (execs who signal safety and set strategy), Lab (a small builder team running experiments), and Crowd (the 80% using AI in daily work). It is built on Ethan Mollick's Leadership, Lab, and Crowd framing from May 2025. Programs that train only the Crowd hit the 20% adoption ceiling and stall. The fix is training all three in sequence.
How is AI workforce training priced?
4 to 8 weeks across one to three tracks. Leadership track: 2 sessions of 90 minutes plus async reading. Lab track: 4 sessions of 2 hours plus hands-on builds. Crowd track: 2 sessions of 60 minutes plus async references. Scoped against team size and track mix, and priced in writing within 48 hours of the discovery call.
How does training fit with the plan and the build?
Training usually runs alongside or after a build. Once a workflow ships, the team that runs it needs the skills to read the evals and spot the next opportunity. The AI Program Diagnostic tells you whether training is the right move at all, or whether the gap is ownership.
Why does AI training usually stall at the 20% adoption ceiling?
The typical program trains only the Crowd. Leadership never signals that AI use is safe to admit, so the secret cyborg pattern takes over: people use AI privately and do not share. The Lab never produces internal examples, so the Crowd has nothing to copy. Training all three in sequence, starting with Leadership, breaks the ceiling.
Who is AI workforce training for?
Companies whose AI direction is set but whose adoption has stalled, usually 20 to 200 people. Execs who need a mental model, and the ICs who need hands-on patterns. It pairs with the Fractional Head of AI Transformation retainer when adoption is the bottleneck.
Can we do just the Crowd track?
You can, but it underperforms. The Leadership track sets the safety signal first. Without it, Crowd training teaches skills people are still nervous to use openly.
How many people can you train at once?
The Crowd track scales to groups of 50. The Lab track stays small, 5 to 10 builders. The Leadership track is the exec layer, usually under 12.
Do we get materials we can reuse?
Yes. Everything goes into the shared knowledge base your team and your AI both read from, or your shared drive. You own the curriculum, and new hires train on it without me.
Is this train-the-trainer?
Optionally. It fits best on the Lab track, so one or two internal leads can run future Crowd sessions.
Where should we aim AI first so people actually use it?
At the work they dread, not the work they are good at. Adoption is a belief problem, and belief comes from relief. When AI takes the annoying work first, people build confidence instead of fear. Ask your most tenured people what eats their week and make them the champions. Low adoption usually comes from fear that AI replaces people, so frame it as help and never start with someone's favorite task.
GAT.
Move your team past the adoption ceiling.
We will place your team on the Leadership, Lab, and Crowd grid, and I will tell you whether training or the plan comes first.
Book 30 minutes with EdwinNot ready to talk? Score your AI program first. Twelve questions, about four minutes, scored breakdown by email.