I recently experimented with Claude and created an AI agent to solve an extremely routine task: the agent was supposed to book a table at a restaurant and tickets to museums in Paris. I asked it to find and book a restaurant in a specific neighborhood, and it did exactly that. What struck me wasn’t that it was able to complete this task. The point is, I had basically handed part of my customer journey over to a machine.
Agents now act on behalf of people, which means that brands will increasingly have to interact with systems capable of finding, evaluating, and purchasing goods and services on behalf of the customer. The concept of “business-to-agent” B2A has already emerged. An agent may not actually make the purchase decision, but it increasingly influences the decision a person ultimately makes.
This creates a new challenge – it’s no longer enough to simply ask whether a customer can find your company online. We also need to ask whether an AI system can understand your company well enough to take it into account when generating recommendations.
AI Is Changing How Customers Discover Brands
The standard formula for digital marketing is relatively simple: companies need to optimize their websites for search engines so that people can find them, and then use marketing tools to convince them to make a purchase. Generative A absolutelyI turns this formula in a different way – instead of providing a list of links for users to explore and choose from on their own, systems like ChatGPT, Claude, and Gemini interpret the question, synthesize information from multiple sources, and provide a specific recommendation.
This changes the very concept of business visibility – you can have hundreds of mentions online and still not show up when a potential customer describes to GPT the criteria for a product that you perfectly match.
AEO starts with understanding how AI perceives your company
The most useful starting point for a company is to conduct a visibility audit of its artificial intelligence systems. Ask various LLM systems the questions that your potential customers would actually ask before choosing a product like yours. Don’t limit this analysis to search queries that include your company name. Ask which companies they would recommend in your category, which products solve a specific problem, how your company compares to competitors, and which option would be best suited for a specific type of customer.
Pay attention to how the company is described, which competitors are mentioned, which sources are cited, and how accurate the information is. You may find that the system understands your category but not your unique selling points, is aware of your product existence but cannot determine which customers it is intended for, or associates your company with outdated positioning.
This leads us to an understanding of what needs to be addressed – for example, making the product positioning much clearer and more structured. Your website and other key digital resources should make it easy to identify and compare the product, use cases, customers, unique features, pricing, integrations, and other relevant information.
This absolutely doesn’t mean that we all have to write exclusively for machines now, but I urge you to ensure that clarity is already part of your business visibility. If you used to focus on storytelling for people, you now also need to provide enough specific information for LLMs.
Your website is just one part of the AI information ecosystem
You have to admit that it’s tempting to reduce an AI visibility strategy to Tier1 and Tier2 publications and other forms of PR, but LLM visibility is much more complex than simply the number of media articles that mention your business.
LLMs can draw on company websites, specialized publications, reviews, forums, social media platforms, directories, comparison sites, interviews, and other third-party sources. Accordingly, it is important for us to build a consistent body of external evidence that will help reinforce the perception of how the company and its product should be viewed.
If a product is described one way on your website, another way on a professional forum, customers use a third description in reviews, and comparison sites like G2 classify it in an entirely different category, it becomes more difficult for AI to reconcile all these signals. The more consistent and reliable the surrounding information is, the easier it is for the system to form a reliable understanding of the brand.
The Next Level of AI Visibility: Brand Understanding
There is an interesting aspect for brands whose value cannot be reduced to the functional characteristics of the product. A good example is the luxury segment. Luxury branding depends largely on heritage, cultural associations, symbolism, understated elegance, and visual language. People can interpret these cues almost intuitively, whereas LLMs typically rely on information that can be identified, matched, and verified.
I’m not saying that brands operating in this space should abandon storytelling and start creating websites that resemble product databases – quite the opposite. The goal is to preserve the brand’s emotional and cultural significance while making enough of its core signals understandable to LLMs.
Conveying values and meaning through AI will likely become an entirely new challenge for companies as AI agents continue to evolve. Algorithms will need assistance in understanding your company and product well enough to determine what information to highlight.
We have started incorporating AI visibility research and reporting into our standard client work alongside traditional PR efforts. A year ago, this might have been viewed as a nice bonus, but now I believe that tracking AI visibility has become yet another fundamental layer in how brands are discovered, evaluated, and chosen.
