The Touchpoint Nobody Budgeted For

Customers now ask a machine what to choose before they reach you. Most marketing budgets still behave as if they don't.

On Wednesday I published an experiment: one question about watching France against Spain in Dubai, asked to ChatGPT twice, logged in and logged out. The two answers did not share a single venue, a three-word challenge flipped a television recommendation, and a major sponsorship never surfaced in a single sentence. [LINK: Wednesday's post]

This piece is about what that night means for the people who sign marketing budgets.

The second brand

The machine's shortlist gets rebuilt continuously from whatever it can read: reviews, listings, articles, structured pages. Reputation that lives in people's memories, and nowhere the machine looks, simply does not make the cut. My own instinct was stale too; I had defaulted to the same bar for the Champions League final in May out of habit. The difference between my memory and the machine's is not accuracy. It is that mine influences eight people on a Tuesday, and the machine's influences every buyer who asks.

That is the second brand: the version of you assembled from whatever the machine could retrieve, described in words nobody at your company wrote. Most companies have never read it.

Millions that never crossed the wall

The television result deserves its own board slide.

The brand the neutral machine recommended first is an official sponsor of the tournament itself. The product images the machine displayed carried the trophy and the partner banner right on the packaging. If ever a sponsorship should surface in an answer, it is this one: a television, recommended for watching the tournament the brand sponsors, during the tournament itself.

It never came up. Not in one answer, in either session, across follow-up questions. The machine recommended the sponsor's television on price and hardware, the way you would recommend a discount. I could not correlate the sponsorship with a single sentence it produced.

Read that as a commercial leader. Money flowed into emotion, awareness, association with the biggest event on earth. At the exact moment a buyer asked the exact question the sponsorship exists to win, the association contributed nothing the machine could repeat. The brand won the recommendation and lost its story. It was chosen as cheap hardware, which is precisely the position a premium sponsorship exists to escape.

The lesson is not that sponsorship is dead. The lesson is that a new touchpoint now sits between every campaign and every buyer, and it only repeats reasons it can retrieve. If the story behind the sponsorship, the technology angle, the credible why, is not written where the machine reads, the millions stop at the door of the conversation where the decision actually happens.

Before you sign the next big commitment, the question to ask is simple: when a customer asks the machine, what will it say about this, and in whose words?

The same person paradox

I decided to build a company on a long walk in the Dubai heat, listening to a podcast about, of all things, billing infrastructure. The founder said something that stuck: machines need the right context to offer the right service.

Look around any office. The same person who asks ChatGPT everything as a consumer, restaurants, televisions, contracts, schools, does nothing as a professional about what that machine says about their own brand. The conversation their customers trust most is the one channel their company neither reads nor briefs.

That gap is not a tooling gap. It is a management gap. Brands maintain a press office for journalists, a sales narrative for buyers, guidelines for agencies. For the interlocutor that now speaks to their customers first, most maintain nothing at all.

Taste, trust, presence

I still have not bought the television, and the reason matters.

The machine gave me specifications, prices, honest trade-offs. What it could not give me is the three things a real recommendation is made of. Taste: a recommender I relate to, with my use case. Trust: someone accountable for the advice. Presence: standing in front of the screen, seeing the picture with my own eyes.

The more I use these systems, the more easily I get answers, and the more carefully I believe them. One challenge in three words flipped a confident "buy this one" into the opposite recommendation. The systems are agreeable by design, and advertising is coming to them. Usage will keep climbing anyway, the way it did with the internet for twenty years. The thing to watch is not usage. It is belief: how much people trust what these systems say, and act on it. The trust layer is moving again, and brands that assume it still lives in their advertising are budgeting for a world that is ending.

Some call this discipline AI visibility. I think the plainer phrasing is more useful: reading what the machine says about you, understanding why it says it, and giving it something truer to say.

What this means on Monday morning

Three suggestions, none of which require a platform.

Ask the machine about your brand the way a real buyer would, logged in and logged out, and screenshot everything. Ask why it did not mention you, because the justification is usually more revealing than the omission. Then put the result next to your current campaign spend and ask whether the story you are paying to tell exists anywhere the machine can find it.

This is the work we started AYAN to do, and the reason we begin every engagement by handing a brand the version of itself it has never read. The brand stays the hero of that story. Our job is to make sure the machine tells it in your words, not its own.

The machine has already met your customer. The only question is who prepared it for the conversation.

Pablo Charron is co-founder and CEO of AYAN, brand management for the AI era.