The climb where four in five stop
On Monday, write two dates on a whiteboard: the day an AI capability became available to you, and the day it ran safely in production. The distance between those two dates is your absorption time, and it is the one number in this chapter that tells you whether your transformation is real. Most boards I speak to cannot produce either date within an hour. That alone is the finding.
Absorption time moves from the budget cycle to weeks when the organisation is built for it. This is part one of two about chapter 17; this part puts the case on the table, the next one is about what you do with it. For now, measure.
Five levels, four climbs
The roadmap in chapter 17 has five levels, and what matters is the four climbs between them. A level is a state. A climb is work, with actions, owners and an exit test that someone has to pass before the next climb begins.
- Deciding: weeks. You choose where AI touches the customer and who owns that choice.
- Connecting and governing: six to twelve months. Data, access and decision rights become one system.
- Redesigning and becoming callable: one to two years. The work itself changes and your organisation becomes something an agent can call.
- Letting the system learn: this one never ends.
Three of the four have an end date. That is what makes them reportable in a quarter.
Why the third climb stops people
Four in five organisations stop at the third climb, and the reason is structural rather than technical. The first two climbs fit inside the organisation you already have. You can decide and you can connect while every department keeps its shape. The third climb asks you to redesign the work and to make your company callable from the outside, which touches roles, budgets and who gets credit.
That is a board decision, taken over a horizon of one to two years, by people who report every ninety days. It stalls because nobody owns the gap between those two clocks.
Three of the four climbs have an end date. The one that pays stops being a project and becomes how the place runs.
The exit test does the governing
Each climb ends with a test that someone passes or fails, which is a stronger instrument than a steering committee opinion. An exit test is written before the climb starts: this capability, in production, for these customers, within this absorption time. It gives the board something to read at the quarter that is not a status colour.
What you ask for
- The two dates, for the most recent capability you put live.
- The name of the person who owns the current climb.
- The exit test, in one sentence, agreed before the work began.
The anchoring kit
Mindset and skillset stay in the organisation when seven things are in place, and training is only the first of them. The anchoring kit in this chapter covers an academy, certification, onboarding, roles and career paths, measurement, reward and perspective. The last two carry the most weight, because people watch what gets rewarded and they watch where this leads for them personally.
A certificate without a career path is a certificate. A career path without a reward is a conversation.
What you do Monday
- Pick the most recent AI capability that became available to you and establish both dates. That number is your baseline.
- Name the climb you are actually on, out of the four, and the person who owns it.
- Write the exit test for that climb in one sentence and read it out at the next board meeting.
- Check which of the seven anchoring elements exist in your company today, and start with reward and perspective.
You now have a baseline and a named owner. The CX+AI Transformation Roadmap sets out the four climbs with their actions, owners and exit tests, so you can plan the next twelve months against something firmer than a wish. View the materials.