The Machine Already Answers for You. The Only Question Is Who Taught It.

Two weeks ago I asked ChatGPT where to watch a football match in Dubai, and it gave me two completely different cities depending on whether I was logged in. This week Deloitte put a number under that small experiment, and the number is larger than most boardrooms have noticed.

More than half of European consumers already use AI to shop. Another 27% say they intend to. In the leading markets, Spain, Poland, Hungary, Italy, adoption is already past 60%. This is not a pilot. Deloitte's own phrasing is blunt: the question for retailers is no longer whether this matters in their market, but how soon.

I want to sit with one finding in particular, because it quietly rearranges how brands should think about the next few years.

## Your customer just split in two

Deloitte's report describes something it calls two customers. There is the human, who still wants control and, above all, trust. And there is the agent, the AI doing the searching, comparing and shortlisting on the human's behalf, which optimizes for information, price and efficiency. They want different things. They respond to different signals. And a brand now has to win both.

Most companies are set up to win exactly one of them. Decades of marketing craft went into persuading the human: the emotion, the packaging, the sponsorship, the shelf. Almost nothing has gone into persuading the machine that now stands between the brand and that human, reading, ranking and recommending before the person forms an opinion.

That is the uncomfortable part. You can win the human's heart and still lose the agent's shortlist, and by the time the human is choosing, the agent has already decided who is on the menu.

## Delegate the search, keep the decision

Here is the finding that should reassure and worry you at the same time.

Deloitte measured what consumers actually hand over. On a seven point scale, people value AI highly for finding and surfacing information, around 5.3. They value it far less for deciding on their behalf, around 2.3. Delegation is high for research and comparison, and it collapses at checkout. The human still makes the final call, still buys through channels they trust: the marketplace, the retailer's own site, the physical store.

Read that carefully and the shape of the next decade appears. The machine is not replacing your customer's judgment. It is deciding what reaches that judgment in the first place. It builds the shortlist; the human picks from it. If your brand is not in the set the machine assembles, the human never gets to prefer you. Their loyalty, your advertising, your reputation in their memory, none of it is ever invoked, because the moment of consideration happened inside an answer they never saw you compete in.

This is why I keep saying that visibility is not a strategy and interpretation is. The machine already answers for your brand. The only real question is who authored that answer, and whether it sounds anything like the brand you think you have.

## Trust is the whole game

The most important line in the Deloitte report is not about capability. It is about trust.

Resistance to agentic commerce, they found, has almost nothing to do with whether the tools are useful. People already believe they are useful. Resistance centers on data, on manipulation, on the fear of losing control. Adoption will depend less on what the agents can do than on whether people trust them enough to let them in.

That should land hard for anyone building in this space, including us. The systems that win will be the ones people believe. And belief is fragile in a very specific way: the first time a machine confidently recommends something that turns out to be wrong about your brand, the trust does not transfer to you, it drains away from the whole channel. A brand described inaccurately by a trusted machine is not a small PR problem. It is a withdrawal from an account you did not know you held.

Which means the work is not to game the machine. It is to make sure that when the machine speaks about you, it is telling the truth, in your words, with the evidence to back it. The market has started calling parts of this discipline generative engine optimization, or GEO. I understand the instinct to file it beside SEO and move on. I think that framing is too small. Optimizing a page is a tactic. Making sure the most trusted interlocutor between you and your customer describes you accurately is a leadership responsibility.

## What this looks like from a Dubai desk

I run commercial operations in a market Deloitte would file under fast adoption, and I see the two customers every week. A shopper in this region will research a purchase through an assistant in Arabic and English, cross a marketplace, a brand site and a physical showroom in a single afternoon, and expect the story to be consistent across all of them. Deloitte calls the agent the connective layer across those channels, the thing that stitches the journey together. The retailers who own that layer shape the whole journey. The ones who ignore it get reduced to fulfilment: someone else wins the customer, and you just ship the box.

There is a hard operational truth underneath this that the report names precisely. Seventy five percent of companies say data is their top barrier to serving this new reality, and only about thirty percent are prioritizing fixing it. The gap between knowing and acting is enormous, and it is exactly the gap where advantage gets built. The brands that move while habits are still forming will define what the machine believes about their category. The brands that wait will inherit whatever the machine decided in their absence.

## The version of you the machine already trusts

I started AYAN because of this exact moment. Not because AI is exciting, but because a second version of every brand now exists inside these systems, assembled from whatever the machine could read, and almost nobody has read their own. We hand a brand that version of itself first. Then we find where it drifts from the truth, and we make sure the machine learns the accurate one, in the brand's own words.

Deloitte's closing line is one I would frame: what retailers decide now, about trust, data and their own AI capability, will determine whether they lead the next decade of commerce or merely supply it. Lead it, or supply it. That is the choice, and it is being made right now, quietly, one answer at a time, whether or not anyone in the building is paying attention.

The machine already answers for you. Someone is going to teach it what to say. The only question worth asking this quarter is whether that someone is you.