Type something one of your own customers might say to you.
A live model will read it in about five seconds. Then I will show you why four people in your own company could each answer that customer correctly and you would still lose the money.
Your words go to the model, it reads them, and that is the end of it. I do not save them and I do not log them. If you choose to share this page, your sentence travels inside the link and nowhere else.
Or start from one of eight examples, in your customers' own words. Yours does not have to be one of them: the box above takes any sector.
Here is what the model made of your words. It read them once, and nobody has touched the result.
All of that came from a live model call on the words you just typed. Everything after this screen was written by me in advance, and none of it is generated.
The model did not come back in time. That happens now and then on the first call of a session, and I would rather tell you than pretend otherwise.
Or look at a saved example from an earlier live run. It is labelled, because it is not live:
Now imagine your own company on the other side of that sentence.
You sell more. You serve more. You be more.
None of that is a technology problem. Every line of it is somebody's job, and in most companies it is four different people's jobs.
“”
Four people in your company would pick that up, and all four of them would answer it correctly. You would still lose the money.
The operation
Somebody has to answer that customer in time, at a cost the business can defend.
I ran that at IAG, with more than 4,000 people across twelve sites. I took nine outsourcing partners down to two, and thirty eight million pounds a year out of the operating cost.
Ask hereWhat does one customer contact cost us, and at what point does the promise stop being true?
The service model
Somebody has to decide what that customer actually meets, and be able to say why it is that and not something else.
I wrote my own service model in 2012 and have used it in six sectors since, all of it before GenAI existed.
Ask hereWhat are we promising customers right now, and who signed it off?
The estate
Somebody has to say whether the CRM, the contact centre platform and the workforce tools you have already paid for can carry the promise you are making, and say no when they cannot.
I know those platforms well enough to tell a board no, and I have had to.
Ask hereCan the CRM, the contact centre platform and the workforce tools we have already paid for actually deliver that promise, and who here is allowed to say that they cannot?
The orchestration
Somebody has to point the AI at the contacts that decide whether a customer comes back, rather than the ones that are cheapest to deflect.
The model you just agreed with is one of two working products I built myself.
Ask hereAre we pointing AI at the contacts that are cheapest to deflect, or at the ones that decide whether the customer comes back?
There are three gaps between those four jobs. Every gap is a handover, both sides of every handover are doing their job properly, and nobody in the company owns the gap itself.
The handover tax
I call that the handover tax. It is the money a company loses in the handovers between the four people who own its customer conversation. It is not lost inside anybody's job. It is lost between them.
At City Link, customers did not know when their delivery was coming, so they rang us, fifty thousand times a day. We started telling them the time instead. The two biggest numbers on this page come from doing that same thing in two industries with nothing in common, and it is what I keep getting hired to do.
Those four questions are yours whether we ever speak or not. Ask all four in the same room, because the answers only tell you anything when you put them next to each other. If they do not fit together, the money is going missing in the gaps between them, and that is the work I do.
If you do want to talk them through, I will take twenty five minutes and I will not pitch you anything. I am not selling an audit and I have no deck.
These numbers come from a three month build that finished in July 2026, for a luxury safari and premium escorted travel brand operating across the Middle East and Africa. It was not a cruise line, and it ran on the contact centre and CRM stack they already had. The client is under a strict NDA, so I cannot name them, and the anonymity is contractual rather than evasive.
35 to 40% more enquiries progressed to quotation · 25 to 28% higher quote to booking conversion · 12 to 15% higher revenue per guest
I am Alex Mead. I am a British citizen, so there is no UK sponsorship to arrange. I am based in Bahrain and I can be on site from week one.