You find out by combining three checks: running test prompts on ChatGPT, Gemini, and Perplexity to see if your practice gets named, asking new patients how they found you during intake, and watching your analytics for visits that arrive without a typical search-engine referral trail. None of these alone proves the full picture, but together they tell you whether AI-driven discovery is happening or whether you're invisible to it.
How to run recommendation checks across engines
A recommendation check means typing the questions a prospective patient would ask into ChatGPT, Gemini, Perplexity, and Google's AI Overviews, then noting whether your allergy and immunology practice appears in the answer. Try phrasings like "best allergist near your city" or "who treats food allergies in your area." Run the same prompts across all four engines, since each pulls from different sources and ranks differently. Do this from a fresh session without prior chat history influencing results, and repeat it monthly, since answers shift as engines update their underlying data. If your practice appears with accurate details, that's a signal you're part of the visible pool. If a competitor comes up instead, or if the engine lists outdated hours, wrong insurance information, or an old address for you, that's useful diagnostic information, not just a disappointment.
It also helps to vary the phrasing the way real patients would. Some searchers ask for "pediatric allergist" specifically, others ask about a symptom like "who can help with chronic hives," and others ask for insurance-specific recommendations. Each variation tests a different slice of how the engine has indexed and understood your practice. Keeping a simple log of which prompts surface your name and which don't gives you a baseline to compare against next month.
Asking new patients how they found you
Intake forms and front-desk scripts are the most direct source of truth for how patients actually discovered your practice, including whether an AI assistant played a role. Add a specific question, such as "How did you hear about us?" with an option for "asked an AI assistant like ChatGPT" alongside the usual choices like referral, insurance directory, or online search. Train front-desk staff to ask this verbally too, since many patients won't remember to fill in a form field accurately.
Patients often describe this experience in their own words rather than naming the technology precisely. Someone might say "I asked my phone" or "I looked it up and it suggested you," which is worth capturing as a note even if it doesn't fit neatly into a dropdown menu. Over time, these notes build a qualitative sense of how often AI-assisted discovery is happening, even before your analytics catch up. This matters especially for a specialty practice like allergy and immunology, where patients often research symptoms extensively before deciding which type of specialist to see, and AI assistants are increasingly part of that research phase.
Signs of AI referral traffic in analytics
AI referral traffic often shows up in web analytics as direct visits or as traffic from domains like chat.openai.com, perplexity.ai, or gemini.google.com, rather than through a traditional search-results click. Because these assistants often summarize an answer and give the user a name and sometimes a link, the resulting visit can look different from a typical Google search click-through. Check your analytics platform's referral source report and filter for these domains specifically, since they're easy to miss if you're only glancing at top-line traffic numbers.
Another pattern worth watching is a rise in visits landing directly on a specific service page, like a page about food allergy testing or asthma treatment, without any referral source at all. This can indicate a patient asked an AI assistant a specific question, received your practice's name along with a summarized answer, and then typed your website address directly or clicked a link that analytics tools categorize as direct traffic. It's not perfect evidence, but paired with the intake question above, a pattern of unexplained direct traffic to specific pages starts to look like AI-driven discovery rather than coincidence.
What to do when engines omit you
When AI engines consistently fail to name your practice, the fix starts with making sure the information these engines draw from is accurate, consistent, and easy to extract. AI assistants tend to pull from sources like your website, business listings, review platforms, and structured data on your site, so gaps or inconsistencies across those sources make it harder for an engine to confidently recommend you. Start by confirming your practice name, address, phone number, hours, and accepted insurance are identical across your website, Google Business Profile, and major directories.
Beyond consistency, make sure your website actually answers the questions patients are asking. If your site doesn't clearly state that you treat pediatric allergies, or that you offer immunotherapy, or which insurance plans you accept, an AI assistant has less material to work with when deciding whether to recommend you for that specific query. Adding clear, plainly written service descriptions and a frequently-asked-questions section that mirrors real patient language gives these engines more accurate, quotable material to draw from. This is also where schema markup, a structured code format that helps search engines and AI systems understand what a webpage is about, can make your practice's details easier to parse correctly rather than guessed at.
Building a simple ongoing check
A sustainable review process means setting a recurring reminder to repeat your AI engine checks, review the intake question data, and scan analytics for AI-referral patterns on a consistent schedule, such as monthly. Treat it the same way you'd treat checking online reviews or your Google Business Profile listing: not a one-time project, but a habit that catches problems before they compound.
Keep a simple record each time you run the check: which prompts you used, which engines named your practice, and any outdated information you spotted. Compare it to the previous month's results to see whether visibility is improving, stable, or slipping. If you notice a competitor consistently appearing where you don't, that's worth investigating further, since it may point to a specific gap, like missing insurance information or a thin service page, that's straightforward to close.
Picture a parent who just moved to your area, searching for a pediatric allergist for their child's peanut allergy. They open an AI assistant instead of scrolling through search results and ask which allergist in town handles pediatric food allergies and takes their insurance. The assistant answers confidently, naming a specific practice across town, describing its focus on pediatric food allergy treatment, and noting it accepts their insurance plan. That parent books the appointment without ever seeing your name, your reviews, or your years of experience, simply because the other practice's information was easier for the AI assistant to find and trust. That's the moment these checks are meant to prevent.