The first version of the internet gave companies a relatively simple job: put your story online. You built a website, wrote an About page, published product information, and gave people somewhere to learn what your company did. If someone wanted to understand your business, much of what they found came directly from you.
Then the internet became participatory. Customers started leaving reviews, journalists and analysts added their own descriptions of companies, competitors published comparisons, and buyers discussed their experiences in forums and communities. The company's website was still there, but it was now one source within a much larger public record that other people were helping create.
AI adds another layer to that evolution.
A buyer no longer has to find and read all of those sources to form their own picture. They can ask an AI system what your company does, who it's for, whether it is reliable, or how it compares with a competitor, and receive an answer assembled from information spread across the web.
This changes the role those sources can play. A G2 review written to help another buyer, a comparison page created to evaluate competing products, an old press article documenting a moment in the company's history, or a community discussion about someone's experience with the product can all be used by AI to generate an answer.
AI is not looking only for your version of the story. It assembles an understanding of your company from the wider public record surrounding it.
Some Sources Are Built to Help Buyers Evaluate You
Third-party review platforms like G2, Capterra, TrustRadius, and Gartner Peer Insights do more than collect customer opinions. Their categories, scores, pros-and-cons sections, and written reviews create structured public descriptions of a product. Taken together, those reviews can give AI a more balanced picture than the company’s own description because they contain evidence of both what customers value and where they encounter problems.
In our work with law firms, for example, we found that AI descriptions draw on legal directories alongside the firm's own website. These sources categorize the firm's practice areas, describe its expertise, and provide external evidence for claims the firm makes about itself.
A comparison page put up by a competitor puts your product head-to-head against its own and other vendors. This page has the agenda of pushing your competitor’s product, but it creates a public interpretation of what your product does, who it is for, and how it differs from alternatives.
These sources were written to help someone understand or evaluate a company or a product. This makes their role as co-authors explicit.
Other Sources Become Co-Authors Without Trying
The more interesting cases are sources that were never created to describe your company.
We saw this in an audit of the managed AI infrastructure market. We ran 141 buyer prompts through Gemini and studied the sources contributing to answers. We found that AI tool companies, who were customers and users of infrastructure platforms, were a source about those platforms.
For example, the AI tool companies had published articles explaining what Replicate was, comparing alternatives to Replicate, and discussing platforms such as Modal and Runpod. They were writing for their own audiences, but in doing so they were also creating public evidence about the infrastructure companies they used.
Imagine a case study on a vendor’s website about Company X using the vendor’s product. Now imagine Company X’s website describing how they use the vendor’s product. The first is the vendor telling a customer story. The second is the customer independently documenting the product in practice, creating another source AI can draw on to understand the vendor.
Customers aren’t the only ones who can play this role. Other sources can become a part of the same public record simply by writing about the company in a different context.
Earned media can become a co-author in different ways. A journalist covering your Series A was reporting a particular moment in your company's history, not trying to maintain an evergreen description of the business. Yet that article can remain part of the public record years later.
Communities like Reddit threads, industry forums, and other community spaces exist because people want to exchange experiences, frustrations, recommendations, and workarounds. They aren't trying to maintain your positioning, but those conversations can still provide evidence about what your product is good at, where it struggles, and how people actually use it.
The common thread is that a source doesn't have to set out to shape your AI persona to become part of it.
How This Changes the Way We Think About AI Visibility
Imagine asking three people to describe the same company: a customer who has used its product for three years, a journalist who covered its last funding round, and a competitor that has compared the product against its own.
You wouldn't expect identical descriptions. The customer can speak to the experience of using the product, the journalist knows the company through the story they reported, and the competitor is describing it in the context of a buying decision. Each contributes a different kind of information.
Your website may provide product facts and technical capabilities. A customer may explain how the product works in practice. A partner may show where it fits into a wider workflow. A comparison site may organize product or pricing differences, while a community discussion may surface implementation problems or real-world experience.
AI encounters all these different perspectives.
This doesn't mean every source will matter equally or that the same sources will appear for every question. It means that your website is only one part of the evidence available when AI needs to understand and explain your company.
The question is no longer only whether the information on your own website is accurate. You also need to understand what the wider public record says about you and who is contributing to it.
A Different Starting Point for AI Visibility
Most conversations about AI visibility begin with the output: how to appear in AI Overviews, how to earn more citations, or how to increase the number of times a brand gets mentioned. Those questions matter, but they focus on changing what AI says before understanding the sources and evidence already shaping those answers.
There is an earlier question worth asking: Who is currently informing AI about our company, and what are they telling it?
The answer may include sources you expected, such as your website, major review platforms, and industry publications. It may also include sources you never considered part of your brand presence, such as a customer's technical blog, a niche comparison site, an old media article, or a community conversation.
Understanding that source environment gives companies a different starting point for AI visibility. Before deciding that the answer is another page, another FAQ, or another piece of optimized content, they can first look at the evidence that already exists around the business and the people and organizations creating it.
Your company is still an author of its story, but it is no longer the only one. The more useful place to begin, then, is not simply with what you want to tell AI next, but with who is already helping AI understand you and what they are teaching it.