Programs · Run it
90-Day Activation Hub
Who runs AI in your organization day to day? Between your AI council and the teams doing the work, somebody has to carry what works to the next team, keep the rules current as the models change, and know what every agent costs. The 90-Day Activation Hub gives that job to two to five of your own people, and spends a quarter building what they run.
Thirteen weeks, three phases. By day 90 you have a charter with a budget line, a roster of every AI pattern and agent in use with a name on each row, a dashboard with real baselines, and somewhere people can ask a question on a Tuesday without booking a meeting. Your people own all four when the ninety days end.
Each quarter adds agents with nobody's name on them, and rules written for models that have since changed.
What one hub carries
Three jobs usually sit in three different places: training on one budget, the council on another, the agent platform on a third. They run on the same handful of people.
-
Learning.
A pattern that works in claims gets carried to billing, written up so a stranger can run it, and counted when somebody who didn't build it starts using it.
-
Governance.
Your AI council sets the guardrails. The hub applies them to the decisions that come up this week, and carries anything with real risk back up.
-
The agent lifecycle.
Every agent and pattern has an owner, a cost and a review date on one roster, and a retirement date when it stops earning its keep.
They belong on one network because they run on the same people. Nobody has to stand it up three times.
What you've built
Your AI leads are active. Teams have been through a workshop and come out with patterns that work. Something is happening, and you can name the teams where it's happening.
The thing most likely holding you back
Your council meets monthly and your builders ship weekly. In between, carrying a pattern to the next team, applying a guardrail to a live decision, and knowing what an agent costs are all jobs, and right now they sit on top of everybody's day job.
How the ninety days run
Thirteen weeks. Lavender weeks need everybody in a room.
Phase 1 · Foundation · days 1 to 30
-
WEEK 1
Kickoff. Team named, board walkthrough, sponsor states the mandate.
-
WEEK 2 · ARENA
Charter workshop. Ninety minutes, everybody.
-
WEEK 3
Knowledge Engine stood up. Governance documents in.
-
WEEK 4
Charter ratified by the AI council. Budget line confirmed.
Phase 2 · Activation · days 31 to 60
-
WEEK 5
Services sized against the hours you actually have.
-
WEEK 6
Office Hours opens. Published widely, same slot weekly.
-
WEEK 7 · ARENA
First Show and Tell. Day 45.
-
WEEK 8
Pattern intake running. First roster entries with owners.
-
WEEK 9
AI leads sync. The network learns what the hub now offers.
Phase 3 · Acceleration · days 61 to 90
-
WEEK 10
First win scoped against one value stream.
-
WEEK 11
Dashboard version one published. Baselines and what you have.
-
WEEK 12
Handoff begins. Every artifact gets a named owner.
-
WEEK 13 · ARENA
Day-90 review. Sponsor and AI council.
The other ten weeks are a weekly call and the work in between. A hub that only moves when a consultant is on the line isn't a hub yet.
Who's actually doing this?
Your sponsor names the team before day 1. Two to five people, which means somebody holds more than one role, and that's expected. What we catch in week zero instead of week six is the same name sitting in three roles with five hours a week behind it.
Melissa · 47 sec
Why Shared AI Responsibility Fails
01
Hub lead
One name. Accountable for the charter and the day-90 review.
Everything below can be shared. This one can't.
02
Enablement
Office Hours, Show and Tell, and the AI leads network.
03
Prototyping and integration
Builds alongside business units. Runs the Knowledge Engine.
04
Measurement
Owns the roster and the dashboard. Chases the numbers.
05
Governance liaison
The line to your AI council. Often part of another role.
Every pattern and agent in use, with a name on it
This is the frame your team opens most, and the one the day-90 review is built on. A pattern moves right as it earns each column, and nothing skips a column.
Melissa · 58 sec
Where Retrospective Action Items Go to Die
-
Heard about it.
Somebody said a team built something. Rumour counts. Write it down before you verify it.
-
Owner named.
A person, not a team. If this is blank, the pattern isn't real yet.
-
On the roster.
An owner, a home a stranger can find, and a thirty-day number.
-
Running elsewhere.
A team that didn't build it is using it. This column is the reason the hub exists.
A pattern with a made-up number is worse than one with a blank, because the blank gets chased. And the fourth column is the one your sponsor should judge you on, which is why it fills last and gets looked at first.
How far past the pilot have most organizations gotten?
Just 16% of large-company C-suite leaders say they've fully designed the roles, processes and operating models to integrate AI into work. Source: Deloitte, October 2025 (C-suite leaders at companies with more than 5,000 employees)
44% of organizations say AI is now scaling across the enterprise, up from 38% a year earlier. The share seeing 5% or more EBIT impact from it is 6%. Source: McKinsey, The State of AI, August 2026 (1,719 respondents, 97 nations)
66% of employees still get limited or no guidance on what to do with the time AI saves them, and 36% say they've had adequate upskilling. Source: BCG, AI at Work, 2026 (about 12,000 respondents across 13 markets)
Start here
Put hours against the spreading
Your patterns work and they stay where they were built, because carrying them is on top of somebody's day job.
Two to five named people have hours for it, a roster that says who owns what, and a review at day 90 that ends in a funding decision.
What you built, and who owns it now
Sixty minutes with your sponsor and your AI council. The charter as ratified. The roster, walking the fourth column and naming the teams. The dashboard, including which numbers you don't have yet. The first win and the claim you're making about it. Three things for next quarter, each with the hours it needs. Then the funding decision, asked for out loud in the meeting.
Melissa · 25 sec
Use AI to Resolve Budget Fights
-
The charter.
Signed, filed where governance documents live, budget line attached.
-
The pattern roster.
Moved into your systems. Somebody keeps it current.
-
The dashboard.
In the spreadsheet leadership already opens. Refresh date agreed.
-
The Knowledge Engine.
Live, curated, with an hour a week against somebody's name.
An artifact without a name stops being true in about six weeks, and nobody notices until they need it.
Before day 1
Two to five people with real hours, named by a sponsor before day 1. A sponsor in weeks 1, 2, 4 and 13. Teams that have already produced AI patterns that work, so the roster has something to hold in week 8.
If your teams haven't built anything yet, start with the Applied AI Workshop. If you have builders and no leads, start with the AI Leads Accelerator.
Proof
From leaders who've learned the model
The Hub is new, and the model underneath it isn't. AI leaders from a global insurer, a Fortune 500 beverage company, a national logistics carrier and a global tech company have been through Running Hyperadaptive Organizations, the class built on the same model. One left having decided to turn a side-of-desk AI program into a funded function with named ownership and decision rights.
A community bank applied the ideas on its own and took a biweekly committee prep task from eight hours to three minutes, then pulled its full rollout forward six months.
- Format
- Thirteen weeks. Three sessions with everybody in a room, a weekly call, and the work in between.
- Your team
- Two to five people, named by your sponsor before day 1.
- Sponsor
- Needed in weeks 1, 2, 4 and 13. Welcome everywhere else.
- Delivery
- Virtual or in person. In-person delivery adds travel and expenses, billed at cost.
- Before you start
- Teams that have already produced AI patterns that work, so the roster has something to hold in week 8.
Designed by Melissa Reeve, delivered by a Hyperadaptive delivery partner.
Questions about running an AI activation hub
- What is an AI activation hub?
- A small named team with assigned hours whose job is spreading what works: running Office Hours, keeping the roster of AI patterns and agents current, connecting teams building the same thing, and holding the line to your AI council.
- We already have an AI council. Does this replace it?
- No. Your council sets the guardrails and the hub is the layer between it and the work, applying them to the decisions that come up this week and escalating anything with real risk.
- Can you run the hub for us?
- No. Your people run it, and by day 90 every artifact has one of their names on it. A hub that only moves when an outside firm is on the call stops moving when they leave.
- Who owns the AI agent lifecycle at the end?
- Each agent and pattern has a named owner in the business that uses it, recorded on the roster. Measurement holds the roster, and the governance liaison carries anything with real risk to your AI council.
- How many hubs does our organization need?
- That question gets raised at day 90 and answered after it. One hub running for a quarter, with a roster and a number, is the evidence that question needs.