Ask ChatGPT, Copilot, or Perplexity where to find a specific kind of service, and you get a short answer: usually two or three named businesses, not a page of links to click through. That is a very different outcome from a traditional search results page, and it changes what a business needs to do to get chosen. Nobody outside these companies has the exact formula, but based on how these systems behave and what they cite, some patterns are consistent enough to plan around.
These Are Not Search Engines Ranking Pages
A traditional search engine ranks a list of pages and lets the customer pick. An AI assistant reads across many sources, then writes a short answer that names a small number of options directly. That means a business is competing to be one of the handful of names the assistant actually decides to say out loud, rather than to appear somewhere on a page one customer may or may not scroll through.
This shift matters because a business that ranks reasonably well in traditional search can still be invisible in an AI answer, if the assistant's underlying sources do not mention that business clearly, consistently, or with enough supporting detail to feel like a safe recommendation.
What These Systems Appear to Lean On
Based on how these assistants respond and what they cite when asked, a few signals show up again and again. Review volume and rating both seem to factor in, though not in isolation, since a business with a handful of reviews rarely gets named over one with a larger, more established review history in the same category. Third-party mentions matter too: roundup articles, comparison pages, and review sites that name a business by name appear to feed into how confidently an assistant recommends it.
Structured data, the behind-the-scenes markup on a page that describes what a service is, its price, and its rating in a format machines can parse directly, also appears to help. A page with clear structured data gives these systems a cleaner source to pull from than a page where the same information is buried in design elements or images.
When a customer asks a specific question, such as whether a plumber turns up on time or how gentle a dentist is, an assistant appears to pull the actual text of your reviews and summarize what real customers said. A business with detailed, specific reviews on Google, Facebook, and Bing gives these systems more to work with than one that only has a star rating next to its name.
Consistency Seems to Matter More Than Volume
A business mentioned the same way across several independent sources, an industry roundup, a review platform, and its own well-structured service pages, appears to get named with more confidence than one that has only ever mentioned itself. This lines up with how these assistants are built: they weigh evidence from multiple angles rather than trusting a single source outright, so agreement across sources looks like it counts for something a single glowing mention does not.
This also means a business cannot shortcut its way in by simply publishing more content about itself. What appears to move the needle is being described accurately and consistently by sources the assistant already trusts, alongside a clean, well-marked-up version of that same information on the business's own site.
Each Assistant Draws From Different Plumbing
Treating the assistants as interchangeable hides a useful detail: they appear to sit on top of different sources. Copilot is built by Microsoft and leans on what Bing has already crawled and cataloged about businesses, which is why a claimed, reviewed Bing Places listing seems to matter more for Copilot answers than for anyone else's. Perplexity shows its citations openly with every answer, so you can see exactly which pages fed a given recommendation, and those citations frequently point at independent comparison and review content rather than brand homepages. ChatGPT behaves differently depending on whether it is answering from what its model already learned or actively browsing the web for the query, and the browsing mode appears to reward the same current, well-marked-up pages a search engine would surface.
The practical takeaway is that visibility is checked per assistant, since being named by one says little about the others. A business can appear reliably in Perplexity answers while staying absent from Copilot's, simply because the underlying sources each system reaches for do not overlap much for that category.
How to Check Where Your Business Stands Today
The baseline exercise costs nothing: open each assistant and ask the questions your ideal customer would ask, such as "who is the best emergency plumber near me" or "good family dentist in my area," and record which businesses get named and, where the assistant shows them, which sources it cited. Run each question a few times, because these systems produce varied answers between runs, and a single response proves very little either way. What you are looking for is the repeated pattern: which competitors come up again and again, and which sources keep feeding those mentions.
Repeat the exercise monthly rather than daily. Answers drift as the systems update, and a monthly snapshot is enough to notice a real shift without drowning in noise. When a competitor keeps getting named, the citation trail usually shows why, and that trail doubles as a to-do list: the roundups they appear in, the review platforms where they stay active, and the way their service pages present information are all things your business can work toward on its own terms.
Practical Steps Based on What These Systems Cite
A few concrete things a business can actually control: keep review platforms current and accurate, since that is one of the clearer signals these systems appear to draw from directly. Add structured data (schema markup) to service and location pages so the underlying facts, service area, hours, rating, are readable directly by machines rather than only visible to a human eye. Where it is honest to do so, get listed in category-specific roundups and comparison content published by independent sites, since third-party mentions appear to carry more weight than anything a business says about itself.
None of these steps guarantee a specific mention in a specific answer, since these systems change frequently and no outside party can see their exact process. What they do is put a business in a stronger position to be cited when the underlying evidence lines up in its favor, which is the most any business can realistically aim for right now.
Why This Is Worth Doing Even Without a Guarantee
The honest position here is that AI search visibility is still a young, fast-changing area, and any claim of a fixed formula should be treated with skepticism. What is not in question is that more customers are starting their research inside an AI chat before they open a search engine at all, so the underlying signals worth strengthening, reviews, structured data, and independent mentions, are worth building regardless of exactly how each assistant weighs them today.