Menu
Close
AI is changing how brands are found, interpreted andrepresented online. The response may have less to do with optimization and moreto do with coherence.
For as long as anyone can remember, brands have had a reasonable amount of control over how they’re seen and how they explain themselves on the internet. Not complete, obviously, but reasonable.
Build the website: Structure the navigation and content hierarchy. Write the About page. Place the right image next to the right headline. Make it look good enough, then plus it up. Create the campaign, dial in the messaging. Publish the case study.
For anyone looking to understand the company, more than likely, they would arrive at somewhere designed for that arrival. And with a reasonable or unreasonable amount of money, anything and everything was possible in terms of visibility.
But AI is renegotiating that arrangement and making a brand good enough isn’t actually good enough.
That’s because people aren’t using ChatGPT, Gemini, Claude and other AI-powered LLMs, like traditional search engines. They’re not looking for a list of five to ten blue links anymore.
They want a point of view, a curated direction, a sense of competence and a conversation. They are asking for who can solve their problem, how a category of products compare and what an organization really does and who has experience in a particular category. Their category.
And after a digital tête-à-tête the answers arrive before they even hit your website.
All of this has spawned a new optimization industry seemingly overnight. Answer Engine Optimization (AEO). Generative Engine Optimization (GEO).Two different names for essentially one big problem: How do we make sure AI can find us and recommend us?
Not an unreasonable question. But I think before anyone oranything can be found, something else needs to come first:
Making sure your brand makes enough sense to beunderstood when you, your website or your marketing is no longer the oneexplaining it.
Institutions have never experienced total control over how their brand is perceived. Customers comment. Employees leave reviews. Competitors make comparisons. Culture does whatever it wants.
Until recently, brands have been able to design many of the environments where understanding happens. To be clear, this is less about the online ecosystem chatter that social media carries and more about digital environments as a whole.
Ferrari’s new EV was panned but managed to sell our completely, largely thanks to wealthy enthusiasts in China looking to flex their electrified swag, despite the social media disdain. But digital environments still hold sway.
A website is one of those environments. Its architecture establishes hierarchy. Copy provides context. Photography creates associations. Case studies provide evidence. Navigation determines what someone encounters next. All those decisions work together to create an intended understanding. AIchanges the conditions.
Ask an answer engine what a company does and it doesn't necessarily experience the carefully constructed journey from homepage to capabilities to case study. It may encounter pieces of the company across its website and other available sources, then assemble those pieces into an answer somewhere else.
The brand hasn't lost control. But some control over thefinal frame has shifted from their control.
The opportunity isn't to control every AI-generated interpretation of the brand. We can't. It's to become much more intentional about the material from which those interpretations are constructed.
And that makes coherence increasingly valuable.
This isn't really an AI problem. Companies have been creating this problem for themselves forever. They grow, acquire businesses, enter new markets, add services, launch products and create divisions. New websites appear. New leadership teams bring new language. Different parts of the organization begin describing similar capabilities differently.
None of those things necessarily creates confusion on its own. The confusion accumulates and proliferates.
Eventually, the organization everyone inside the company understands is not quite the organization everyone outside it sees.
The problem becomes particularly obvious in complex businesses. A group can own five companies that make perfect strategic sense together while presenting five completely different explanations of what the larger organization actually does.
A restaurant group can operate a collection of individually strong concepts without making the value of the group itself clear. A university can have an extraordinary collection of schools, programs and initiatives that somehow add up to a surprisingly vague institutional identity.
The underlying business may be perfectly coherent but itsrepresentation isn't.
In Ways of Seeing, John Berger explored how reproduction changes our relationship with images. A painting once encountered in a particular place could suddenly appear on television, in a book, beside another image or accompanied by different words. The image remained, but the context around it changed, and context changes how something is understood. I think that brands have lived with versions of this problem for a long time.
A logo appears somewhere the designer never intended. An advertisement gets separated from the campaign around it. A quote gets pulled from an interview. A product becomes part of a subculture the company never anticipated. Social media accelerated all of it.
AI introduces another version of the same problem. Information can leave the experience it was designed for and become source material for an answer assembled elsewhere.
This doesn't make intentional brand experiences obsolete. It makes the ability of the brand to remain understandable outside those experiences more important.
That's a different standard for coherence.
The obvious response is standardization. Make every description match. Use identical terminology everywhere. Structure every page predictably. Remove ambiguity. Repeat the same positioning until there is absolutely no possibility of misunderstanding.
That might make a company easier to parse. It could also make it less of a brand. Nearly a century ago, Japanese artist and color theorist Sanzo Wada confronted a version of this tension from a completely different direction.
Japan was modernizing and participating in an increasingly global industrial economy. Color was no longer only something experienced locally through art, nature, textiles and traditional materials. It increasingly needed to be communicated, specified and reproduced across commercial systems.
That created a practical problem no one had encountered before: color needed a shared language.
Western color systems offered scientific methods for classification and standardization. But traditional Japanese understandings of color carried relationships to seasons, materials, nature and emotion that could not be fully expressed by reducing individual colors to mathematical positions on a chart.
Wada's response was not to reject systems or systems thinking. It was to help shape them. But his work also preserved something systems can easily lose: relationship and legibility.
His color combinations weren't simply catalogs of isolated shades. They demonstrated how colors behaved together. The system created enough structure for color to travel while preserving the perceptual relationships that gave those colors life.
Which is precisely how institutions should be thinking about their brands.
The challenge isn't simply making meaning consistent. It's making meaning capable of traveling without losing what made it meaningful.
A brand isn't any single expression of itself. A logo isn't the brand. Neither is a website, campaign, photograph, service page,restaurant, product or piece of thought leadership. Each contributes something,and much of the meaning exists in the relationships between them.
This is why coherence and uniformity are not the same thing.
A restaurant group can operate concepts with entirely different names, interiors, menus, voices and audiences while still creating a coherent portfolio. An acquired construction company can retain decades of history and local equity while becoming an understandable part of a larger enterprise. A global consumer brand can move between products, athletes,campaigns, cultures and generations without making every expression look identical.
The pieces can change because something underneath them holds. That's what allows the brand to travel. And as AI becomes another intermediary between brands and people, the ability to travel matters more.
Information will increasingly appear outside the environment it was originally designed for. It will be retrieved, summarized, compared and combined with information from other sources. The goal can't simply be to maket hat information machine-readable. It still has to feel like something.
We recently saw a small version of this while working with a highly specialized marine and industrial contractor.
The company operates in a technical world involving different vessel types, regulatory standards, electrical systems and customers who care considerably more about compliance and reliability than clever marketing.
We weren't trying to optimize the company for AI. We were trying to make a complicated business easier to understand.
The site was structured around direct questions and factual answers. Technical claims became more specific. Terminology became more consistent. Different pages reinforced the same underlying understanding of what the company did without simply repeating one another.
Later, the company began appearing in AI-generated answers for relevant category searches. That doesn't prove we cracked an algorithm. We didn't. But it raised an interesting possibility.
The same decisions that made the brand more legible to people may also have made the company easier for machines to interpret.
This is where I think the conversation about AEO and GEO becomes much more interesting than another set of optimization tactics.
Legibility now has two jobs. The first is informational. A person or machine should be able to understand what the organization is, what it does, how its parts relate and why its claims are credible.
The second is emotional. The brand still needs to create recognition, trust, familiarity, belonging, agency, meaning or whatever combination of human responses the strategy calls for.
Solving the first at the expense of the second gives us an extraordinarily organized company nobody cares about. Solving the second without the first gives us an interesting brand nobody can explain.
There are real technical questions surrounding AI discovery. Search fundamentals still matter. Structure matters. Authority matters. Specificity matters. The mechanics will continue evolving as the platforms evolve.
But none of them removes the more fundamental brand question. What are we asking these systems to understand?
If the company describes itself differently depending on where you encounter it, if its architecture is unclear, if its claims are generic, if its businesses have no understandable relationship to one another,the problem exists before an AI system ever encounters it.
That's why I think coherence may become one of the more important brand assets of the AI era.
Not rigid consistency. Not endless repetition. Not sacrificing personality so information can be perfectly categorized. Just enough structure that the brand remains understandable as it moves between contexts. Enough consistency that it remains recognizable even when the expressions change. Enough clarity that the pieces continue to add up to the whole.
Wada was trying to solve a very different problem in a very different time. But there is something useful in the instinct behind his response to modernization: Create a system strong enough to travel without standardizing away the thing that made it worth carrying.
Brands now face their own version of that challenge. We're going to have less control over some of the frames through which people understand us. That’s not to say we've lost our ability to influence what they see. It just means we need to become much more intentional about what survives the journey.
Written by: Jeff Skatzka
