As AI becomes an intermediary between brands and customers, businesses need to be understood by machines without becoming forgettable to people.
AI is changing how people find brands, and it’s beginning to change how they experience them. Search engine AI overviews, LLM recommendations and emerging purchasing agents increasingly sit between a customer’s need and a company’s carefully designed brand experience.
Before someone visits your website, sees your campaign or walks into your store, AI may already have summarized your brand, compared it with competitors and decided whether it belongs in the consideration set at all. That creates a new challenge for businesses.
Brands now have to succeed in two systems at once: they need to be legible to machines and desirable to people. The temptation is to treat the first of those as a new optimization problem. Improve AI visibility. Generate more content. Track mentions. Understand citations. Find ways into the answer.
But our latest research into the pressures facing CMOs suggests a tension. At the same time that organizations are directing more investment toward AI, many are pulling investment away from the very foundations that influence how AI understands and represents a brand: UX, owned digital experiences, thought leadership, content and the wider customer ecosystem.
In other words, businesses risk chasing the algorithm while weakening the infrastructure that feeds it. That matters because AI does not only need to know that a brand exists. Increasingly, it needs to understand what the brand stands for, why it is different, whether its promises can be trusted and—perhaps most importantly—who it is for.
The mattress category offers a useful illustration
Casper helped make the mattress showroom disappear. No fluorescent lighting. No three-minute test nap with a salesperson standing six feet away.
Casper made the category simpler, friendlier and infinitely more memorable. So, when we asked AI for the best mattress for a side sleeper with back pain, we expected Casper to be all over the answer.
It wasn’t. Saatva and Helix appeared more naturally around specific needs and use cases. According to Mike Ensing, CEO of the AI Brand Intelligence company Revere, Saatva and Helix are the visibility leaders across the major LLMs with them dominating positioning for mattresses for side sleepers with back pain, Casper and others being way down the list.
AI knew Casper. But it knew what to do with Saatva. When we first looked at the category, Casper was still largely selling one big idea: a better way to buy a mattress. Saatva was selling answers. Side sleeper? Start here. Back pain? Look here. Sleep hot? This might be yours.
Casper had stronger name recognition. Saatva had a better dating profile. Its portfolio and the evidence surrounding it gave AI a clearer bridge from a particular person to a particular product. Casper was more famous, but Saatva was easier to match.
For years, digital strategy has assumed that people want access to more options: more products, more comparisons, more reviews, more information, and more tabs left open until someone finally gives up and buys the thing they saw first.
But people do not really want more choice. They want to feel confident about a choice.
Choice is a tax. AI has volunteered to pay it.
People are already using AI to learn, compare, narrow the field and make recommendations. The next step is not hard to imagine. As trust grows, more people will allow an agent to manage parts of the choice itself, within whatever rules and limits they set.
That changes what visibility means. Search made visibility a ranking problem. You wanted to appear near the top of the list. AI turns visibility into an inclusion problem. Either the brand becomes part of the answer, or it may never become part of the decision.
Search refers, but AI matchmakes.
An agent interprets the request, weighs the evidence, considers the person and decides which brands belong in the conversation. It is becoming a new kind of experience stakeholder, with its own system of judgment and an increasingly important role between the customer and the company.
AI’s problem is not, “Have I heard of Casper?” It’s “Would I put this particular person on one?” Finding a brand and knowing who it belongs with are two very different acts.
Your portfolio is becoming a customer interface
Your brand is the face of your business strategy. It tells the world what the company believes, where it intends to compete, what kind of value it wants to create and why anyone should care.
But the portfolio is where that strategy becomes a set of actual choices. Which customer gets which offer? Which need does each product solve? When is one option better than another? What does the business understand about people that its competitors do not?
For a human customer, a strong brand can create meaning, memory and desire. For an AI agent, the portfolio creates matchability.
This makes portfolio architecture much more than an internal product-management exercise. It is becoming a customer interface. It is one of the primary ways a business makes its customer strategy visible to the systems increasingly responsible for interpreting customer intent.
A vague portfolio forces AI to infer. A clear portfolio gives AI a decision architecture. And when an agent is deciding what belongs with whom, clarity has commercial value.
From customer centricity to LLM-to-user centricity
Traditional customer centricity asks a company to understand the customer and organize around their needs. The trouble is that much of this understanding remains trapped inside the company. It lives in a segmentation deck, a research repository, a product brief or the memory of a few smart people who keep getting invited to all the important meetings.
An AI agent cannot reward customer centricity it cannot see. I’ve started calling this LLM-to-user centricity. It is not the most graceful phrase I’ve ever coined, but it names an increasingly important job.
LLM-to-user centricity is the discipline of structuring a business, brand and portfolio so an AI system can connect a person’s needs, preferences and circumstances to the right offer, then explain why the fit makes sense.
If the brand is the face of the business strategy, the portfolio is the key operating layer of LLM-to-user centricity. It is where customer understanding stops being an internal insight and starts becoming a visible choice.
This is not about writing robotic copy for machines. It’s not about flooding the internet with synthetic content. And it is not a new name for search optimization.
It is about making the actual logic of the business legible: the products, the audiences, the needs, the differences, the evidence and the reasons one offer belongs with one person rather than another.
Customer centricity says, “We know our customers.”
LLM-to-user centricity asks, “Can the agent acting on their behalf tell?”
The rise of the upper-middle visibility darling
The mattress example suggests another possibility. AI may create an unusual advantage for a certain kind of upper-middle brand.
These brands are large enough to have authority, reviews, customer evidence and third-party credibility. They do not feel obscure or risky. But they are often focused enough to be specific. They may have clearer use cases, sharper audience definitions and more explicit reasons to choose one product over another.
The largest brand often owns the category noun. The upper-middle brand may own the words that come after “for.”
Running shoes for a first-time marathoner.
A moisturizer for sensitive skin.
A CRM for a startup. A mattress for a side sleeper with back pain.
In our early category work, this pattern appeared beyond mattresses. Nike had the culture and category equity, while Hoka, Brooks and On appeared more naturally around specific running needs. Salesforce owned CRM mindshare, while HubSpot was easier to connect to startups and small businesses. Prestige skincare brands carried human affection, while CeraVe, The Ordinary and La Roche-Posay benefited from clearer clinical and use-case evidence.
This is a hypothesis, not yet a law. Small brands can be highly specific but lack enough evidence to feel safe. Large brands can have enormous authority but become so broad that AI struggles to place them.
The upper middle can sit in a powerful position. Trusted enough to recommend, and specific enough to match.
Four trust signals. Two questions.
Our AI Visibility model looks at this emerging decision environment through four trust signals, organized around two questions.
First: Can AI find you?
Legibility[SL1] [SL2] asks whether AI understands your brand, your positioning and the architecture of what you offer. Can it digest the business well enough to explain it accurately?
Distinctiveness asks whether AI can tell you apart from the category. Do you look and sound like yourself, or have years of category convention turned you into a slightly rearranged version of everyone else?
Then comes the more consequential question: Will AI choose you?
Reliable asks whether AI trusts what you deliver enough to recommend it. It looks for consistency between the promise, the experience, the evidence, the campaign, the community conversation and what other credible sources say.
Compatible asks whether it is clear who you are for, and whether AI knows who that is. Can it connect a particular customer, need or circumstance to your brand and explain why the fit makes sense?
Together, Legible, Distinct, Reliable and Compatible create a trust engine. Legible and Distinct help AI find and understand you. Reliable and Compatible give it the confidence to choose you.
Compatibility may be the sleeper signal.
A company can improve its content, metadata and technical discoverability and still have a compatibility problem. If the portfolio does not make clear who gets what and why, AI is left to assemble the customer logic on its own.
Search optimization can make a product easier to find. It cannot decide whom the business intends to serve. That is why AI visibility is not merely a communications challenge. It is a diagnostic of the business strategy itself.
When AI repeatedly fails to recommend a brand for a high-value need, the problem may sit in the positioning, the portfolio, the naming, the product experience or the evidence surrounding the promise.
The answer might be better content, but it might also be a better-organized business.
The visibility opportunity is larger than mentions
Most companies will initially treat AI visibility as a measurement problem.
How often are we mentioned? Where do we rank? Which sources are being cited? Those questions matter, but they are only the beginning. A mention is not the same as a recommendation, and a recommendation is not necessarily the same as a good match.
The larger opportunity is to understand what AI believes about the brand, why it believes it and what would need to change for the brand to become a more natural recommendation.
Executed properly, the entire marketing and experience system becomes a trust engine. The brand strategy, portfolio, content, product experience, reviews, community conversation, campaigns and third-party coverage all reinforce a coherent answer to four questions:
What are you?
Why are you different?
Can you be trusted?
Who are you for?
That coherence is useful to machines because it is useful to people. A clearer portfolio helps a customer choose. A sharper value proposition helps a customer understand. Better evidence helps a customer believe. More explicit compatibility helps a customer feel that the product was designed for someone like them.
Casper helped make the mattress showroom disappear.
AI may now make the shortlist disappear.
In that world, being known is useful. Being matchable is what gets you chosen.
