Proof Positive
Un regard bimensuel sur l’advocacy
client à l’ère de l’IA.
Vingt-cinq ans de pilotage de programmes de références et d’ambassadeurs clients, condensés en une courte note toutes les deux semaines. Ni opinions à l’emporte-pièce, ni catastrophisme sur l’IA : simplement ce qui fonctionne, ce qui ne fonctionne pas et les prochaines étapes.
Derniers numéros
The nine-week queue
Review cycles are where customer stories go to die. Nobody in the cycle took nine weeks; five or six people each took a reasonable week, one after another, and the story spent most of that time waiting in an inbox. Red tape is real and so is urgency, but most of the delay is neither; it is a queue nobody can see, and it can be fixed without short-changing the advocate or the teams who pay for the work.
We don't add, we average
Show people a perfect set of 24 plates and they will pay more for it than for a set of 40 with a few broken. We judge an offer by its average, not its total, and customer stories are no exception: every weaker proof point you add pulls the best one down. And the reader that now goes first, an AI assistant, will lift a sentence or two; with eleven proof points to choose from, you do not get to say which.
Keep the best, speed up the rest
Half of consumers say they would rather buy from brands that avoid generative AI, and I understand why. But a no-AI rule protects a nine-week queue and a four-story year. Keep the interview, the advocate's experience and the sign-off exactly as they were; speed up everything the advocate never sees; and spend what you save on the stories nobody ever got round to.
Showing my working
Last issue I argued that a story is no longer its own proof; it needs the working shown underneath it. This week I am asking for more trust than I have asked for in this publication so far, so here is my working: what I built, the three things I got wrong building it, and what my own guarantee does not cover.
Seeing is no longer believing
In 2023 the dictionary's word of the year was authentic. In 2025 it was slop. Two words, two years, one story; and it makes the work customer advocacy people do more valuable than it has ever been, not less.
Big logo, small story
Everybody wants the famous name on the slide. But your largest customers by revenue are rarely the ones who influence the most revenue, and the companies who complain loudest about enterprises refusing to advocate usually cannot say who signs off a reference request at their own company.
The interview is (nearly) everything
Ninety percent of their revenue came from ten customers. Not ten percent; ten. So which of those ten do you hand to an online form? Drafting stopped being the expensive part, which means the interview is the whole game, and at the top of your customer list it is not an input at all. It is the relationship.
Writing was never the bottleneck
The draft is finished on day two. The story publishes in week nine. Everything that happens in between is the actual job; and it is the part nobody has automated, because it was never a writing problem in the first place.
Not all stories are equal
Some customer stories will never be great, whatever you do to them; others are far bigger than they first appear. The skill is matching your investment to the story, and being honest that investment means human time, which means money.
Citizen, marketer, editor
For thirty years, the hardest part of producing a customer story was beginning one. That has changed. The work has moved to everyone who touches the story; we have become a profession of Citizen Editors, whether we have recognised it yet or not.