Find your zero
The demo always works. The assistant answers three questions in a row, the tone is warm, the room nods, and somebody says we should put this live before the quarter closes. I have sat in that room more than once. The energy is real and the model is genuinely good.
Then someone asks what happens when the assistant hands the customer over to a human. The room goes quiet. That silence is the most useful thing in the whole session, because everything worth fixing lives in it.
The moment after the handover
The handover is where your organisation gets tested, not the model. The assistant passes the conversation to a team that works from a different script, in a different system, with a different target on the wall. The customer explains the situation again. Within four messages the goodwill the assistant built is gone.
Walk backwards from that moment and you find the same four things every time:
- No single owner for the journey, only owners for the parts of it.
- Context that the assistant has and the human team cannot see.
- Decision rights that live two levels above the person in the conversation.
- A target that rewards closing the contact, not solving the problem.
None of those is an AI problem. All four were there last year. The assistant simply made them visible on a Tuesday morning in front of the board.
AI multiplies what is already there
AI behaves like a multiplier on the process you point it at. A journey that scores a four becomes a very loud four, at volume, in front of more customers than before. A process without an owner becomes an unowned process running thousands of times a day. A multiplier on zero returns zero, and no amount of model quality changes that arithmetic.
AI will not transform your organisation. It will expose it.
That is why the honest first move is to find your zero. The one journey, handover or decision where your score is nothing, and where every bit of AI investment you make is being multiplied by it.
Ten shifts, none of them technical
The book describes ten shifts in thinking, and not one of them is about technology. They sit in three places, which is useful when you are trying to work out who needs to be in the room.
Where the shifts sit
- Organisation: who owns the journey, who holds the data, who may decide what without asking.
- Customer: what the customer now expects from a company that answers in seconds.
- Leadership: what a director measures, funds and protects when the easy wins are gone.
Your tooling budget touches none of these three. Your operating model touches all of them.
Maturity is climbed
You cannot buy maturity and you cannot wait for relevance to arrive. Climbing is the only route available, and it goes one step at a time. Pick one journey. Give it one owner with a name. Move one decision down to the person who speaks to the customer. Measure the handover rather than the demo.
That is slower than a procurement cycle and it holds. Teams that do this get a second step for free, because the first one teaches them where their next zero sits.
What you do on Monday
- Replay five real conversations from the point where your assistant hands over, and write down what the customer had to repeat.
- Name one owner for one end to end journey, in writing, this week.
- Score that journey honestly on a scale of one to ten, then ask what AI multiplies it by.
- Move one decision down a level and watch what happens to handling time.
- Book thirty minutes with your data lead on what the human team can see and what the assistant can see.
If you want a structured way to locate your zero before you scale anything, the CX+AI Maturity Check walks you through the organisation, customer and leadership shifts in one session. It gives you a score per area and a first step that belongs to a person, not a project. View the materials