AXHUB5 cards

A leader's 90 days

The real work starts the day after kickoff.

A published 90-day adoption script, organized into five stretches.
Listen and choose, build with real data, embed (the make-or-break stretch), prune, and fit to scale.

Organized from Treetop's public playbook (consulting material) — performance figures are marked as targets

A leader's 90 days1 / 5

Weeks 1-2 — listen and choose

Starting to build right away is the most common failure path.
The first 2 weeks are for listening.

Interview three or four people per role for 30 minutes each and 8-12 candidate workflows emerge.
Prioritize by "frequent, tedious, and rule-based," in that order.

Two interview questions are enough "Among your weekly repetitive tasks, which wasted time hurts most?" "If you handed that task to a new hire, how would you explain it?"

Source 90-day playbook discover stage — "skipping discovery" is the most common collapse point

A leader's 90 days2 / 5

Weeks 3-6 — build with real data

Build the top two or three with actual work data.
No demo samples — use last month's real inquiries and real documents.

Don't wait for completion; ship a real output first, within this stretch.
Users' trust comes not from a demo but from the first real result.

The output of this stretch 2-3 working workflows · a shared team prompt collection · a few outputs that actually shipped

Source 90-day playbook build stage — "build with real data" is non-negotiable

A leader's 90 days3 / 5

Weeks 7-9 — embed, the make-or-break stretch

The stretch the playbook nails as "where most fail."
There are three things to do.

Small-group training by role, three or four times, instead of a generic lecture.
One or two champions (heavy users) per department.
The team's AI rules, written down.

A developer can't design the training alone.
Only someone who's done the work knows what's judgment and what's task.

Example table of contents for the rules doc 1) Data you may / may not put in 2) Output review criteria (who, what) 3) Where to keep good prompts

Source 90-day playbook embed stage · "training design needs field experience" (practitioner view)

A leader's 90 days4 / 5

Weeks 10-12 — prune

Open three weeks of real usage data.
What's used daily and what's abandoned split clearly.

The knack is not to strain to save the abandoned.
Even if the effort of building it stings, clear it out and put your weight behind what works.

Three columns for the audit Per workflow — uses in 30 days / number of users / decision (keep, reinforce, clear out)

Source 90-day playbook optimize stage — watch for "defending a failed workflow"

A leader's 90 days5 / 5

Scale it down or up to fit

For a team under 10, even 90 days is long.
Compress to 45 days, and two workflows are enough.
For 50-150 people, do the reverse and stretch to 180 days.

One budget reference —
the organizations that succeeded spent 1 on tech and 9 on training and embedding.

Source 90-day playbook scale variants · 1 tech cost : 9 change management (Accenture Federal, Axios)

AXHUBclosing

What remains is a habit,
more than a number

The playbook's 90-day targets are numbers like "60% daily usage, 60-70% cut in prep time."

But surer than the targets —
a team that spends 90 days in this order is left with a habit, not just a tool.

See the source material on axhub.net

Sources: Treetop 90-Day AI Rollout Playbook (consultancy public material — figures are targets) · Accenture Federal (Axios) · practitioner view

AXHub card No.6 — the playbook figures are the consultancy's targets and experience, not verified research. The example boxes are suggestions to follow.