Marketing has always started from a relatively comfortable premise: somewhere along the way there was a person making the decision.
We could argue about algorithms, programmatic media, search engines and recommendation systems, but in the end we pictured someone looking at alternatives, comparing attributes, developing a preference and choosing.
That premise is beginning to change.
Artificial intelligence systems are already taking over ever larger parts of the discovery and purchase process: comparing products, interpreting specifications, assessing compatibility, hunting for conditions and, in some cases, going all the way to the transaction. Google, for instance, has been building infrastructure specifically for agentic commerce, including protocols that let systems search, interact with retailers and carry out steps of a purchase on people’s behalf.
The change looks technological, but there is a far more interesting marketing question behind it.
A brand may now have to convince two audiences at the same time: the person who wants something, and the system that helps that person decide which something to choose.
A machine isn’t impressed the same way a person is.
For decades we learned to build perception using elements that go far beyond a product’s objective characteristics. Story, desire, aesthetics, reputation, identification, cultural context and emotion are part of what we call a brand precisely because human beings don’t choose by the rational sum of attributes alone.
An AI agent introduces another layer into that equation.
If someone says “find me the best option considering durability, technical support, delivery time and customer reviews”, much of the consideration process can happen without that person visiting ten websites, watching five brand films or scrolling through a page of results.
The machine will look for evidence it can interpret.
Which means a company can be very good at looking desirable to human beings and, at the same time, surprisingly hard for systems to understand.
Accenture already describes this movement as the emergence of two audiences in commerce, the consumer and the agent acting on their behalf, arguing that brands will need to be not only desirable, but also verifiable and understandable to these new interfaces.
That may be a bigger shift than it seems.
For a long time, marketing worked to reduce the importance of objective comparison.
A strong brand pulls off something extraordinary: it turns products that could be compared on specifications alone into choices loaded with meaning.
That won’t disappear because an artificial intelligence showed up. People will keep wanting things for complex, emotional and even contradictory reasons, as they always have.
But part of the journey may be filtered before it ever reaches them.
If an agent rules a brand out because it can’t verify a particular piece of information, because it finds inconsistent data, because it doesn’t quite understand a product difference, or because other companies offer clearer evidence, the battle for perception may never even take place.
Which raises an interesting problem: how do you keep building desire without becoming illegible to machines?
The answer probably isn’t turning brands into gigantic spec sheets, just as it isn’t starting to write every piece of content “for the AI”.
It’s improving the quality of whatever holds the promise up.
Promising will remain the job of communication. Proving may become more important.
Companies have grown used to operating for many years with a certain distance between narrative and information.
The campaign says one thing, the corporate website explains another, the product page holds different information, distributors use old descriptions, news stories quote outdated numbers, and internal presentations keep data that never reached the public channels.
A human being can live with that mess, up to a point. They interpret context, fill in gaps and sometimes simply believe the most convincing version.
Systems that have to compare alternatives tend to demand something else: consistent, structured, available and verifiable information.
That can even change the relationship between marketing, corporate communication, product, technology and data. Information once treated as an operational detail starts to play a direct role in whether the brand can be found and recommended at all.
The work stops being only about telling a story well; it also involves making sure there is an information structure capable of holding that story up when someone, or something, decides to check it.
This doesn’t mean starting to do marketing for robots.
That would be an easy conclusion and probably a wrong one.
If every company starts optimizing its brand purely for the logical criteria of machines, we risk creating a market that is extremely comprehensible and profoundly dull.
Brands exist because people attach meaning to things, and meaning doesn’t come from verifiable attributes alone.
The interesting challenge lies precisely in holding both dimensions.
Being emotionally relevant to whoever chooses, but clear enough for whoever helps them choose. Having personality without sacrificing precision, building desire without hiding information, developing strong narratives while keeping a factual base coherent enough for different systems to understand it.
That may require a convergence that hasn’t really happened yet in many companies: brand, content, product, SEO, technology and data no longer operating as separate territories.
The new question may not be only “what does the consumer think of our brand?”.
It may also be necessary to start asking what the systems that mediate decisions are able to understand about it.
Do they know exactly what the company offers?
Can they identify what genuinely sets its products apart?
Do they find consistent information?
Do they understand who a given solution is for, and who it isn’t for?
Do they recognize enough evidence to justify a recommendation?
These questions don’t replace any of the traditional marketing questions, but they add a new layer.
For decades, brands learned to compete for space in people’s minds. Then they learned to compete for position in search engines, feeds and platforms.
Now a new place may be emerging where that competition will happen: the reasoning of systems that help people decide.
And being present there will probably depend less on repeating that a brand is the best, and more on making clear why someone, human or not, would have reasons to choose it.
