AXHUB LECTURE · INTERMEDIATE L6
A leader's 90 days, 30 minutes today
— build a one-page plan
The 90-day script for team adoption is in card No.6. This lecture is a practice for mapping that script onto your team. Thirty minutes from now you'll have a one-page plan with one problem, a first-4-weeks schedule, three rules, the people to meet first, and a review cadence. This one page comes before the tool-purchase request.
1Pick one problem (5 min)
"Let's do AI too" is not a plan. Organizations that succeeded started from one annoying problem.
The criteria for a candidate are the same as for individuals — clear input, repeated weekly, and it shows when it falls behind.
Goal in 90 days: ____
e.g. "first response to customer inquiries averages half a day → within 1 hour"
If no number comes to mind, that's your first discovery — write it as measurement item No.1 for the week-1 interviews.
Check: does the sentence include the task name and a number (current → target)?
2Plan the first 4 weeks, not the 90 (7 min)
The full structure of the 90-day script is Discover (weeks 1–2) → Build (3–6) → Embed (7–9) → Optimize (10–12).
Today, make just the first 4 weeks concrete. The rest sharpens on its own once the front works.
Week 2: sort candidates, pick the first workflow
Week 3: build and test with real material (no fake practice data)
Week 4: 15-min team share + start the second workflow
The pace benchmark is "one workflow per week." For a team under 10, you can compress this script to 45 days.
Check: does the week-3 box still say "real material" — building on fake data is one of the six failure points.
3Write the three rules first (5 min)
Absence of a governance document is also among the six failure points. Not a dozens-of-pages doc — three lines get you started.
What data is allowed. Review and accountability. Records.
(2) Outgoing outputs are checked by an owner, and accountability rests with whoever checked
(3) Leave the instructions that worked and the accidents in the team document
Just rewrite this example in your team's language. Card No.8 has the background for these three lines.
Check: are the three lines sentences everyone on the team can understand (no jargon)?
4Go to where it will get stuck, first (6 min)
The top source of adoption resistance wasn't the front line but support teams like legal, HR, risk, and security (frequency 35%, front line 23%).
The other half of the same analysis — involve them early and they often turn into allies.
So before the plan is finished, go to them first.
Our team wants to try attaching AI to the ____ task.
Before we start, we'd like to align on the standards to follow.
We'll bring our draft three-line rule, so could you spare 30 minutes?
The point is not "please approve" but "let's set the standards together." Being informed and being invited draw different reactions.
Check: does the note read as a request to collaborate, not a request for approval?
5Create a place to measure (7 min)
In the 90-day script, success or failure is decided not in weeks 3–6 of attaching the tool, but in the embed phase after.
What creates embedding isn't an event but a cadence — 15 minutes weekly, same weekday, one metric.
And a retirement principle. Don't try to preserve a workflow that isn't used — drop it. Separating what's used daily from what's abandoned, via a real-use audit, is the work of the optimize phase.
Three fixed agenda items: (1) this week's metric ( ____ ) (2) one instruction that worked (3) what to drop / what's stuck
One-line metric: "of this week's ____, the share that went out without human edits ____%"
Fixing the agenda at three keeps the meeting from dragging. Only one metric — past two, you stop measuring.
Check: did you gather today's five items (problem, 4 weeks, three lines, note, invite) into one document — that's the plan.
When it doesn't work
First lecture to recommend to your team → L1. Start with one task you repeated today · Full contents
Sources: Treetop 90-day rollout playbook (week-by-week script, six failure points, 45-day compression, success-criteria examples, public material) · Stanford Enterprise AI Playbook (resistance source 35% vs 23%, 77% of the hardest part is beyond technology) · Asrify (one-per-week pace, positioning message, noted as secondary). Full sources are in the axhub.net case library. The practice templates are example lines to follow along.