What does my AI Visibility score actually mean?

Sound familiar?

"My AI visibility score has gone down, not up." "We've been publishing and the number hasn't moved." "Why was I named last time and not this time?" "Why has my score changed when nothing on my website has?" These are fair questions, and the answer starts with what the score actually measures.

What's happening

Your AI Visibility score shows how often AI assistants recommend your clinic when a patient asks for help. We put the same patient questions to ChatGPT, Claude, Perplexity and Google's AI Overviews. Each one is a short conversation that ends with the patient asking who they should see. That is 64 answers in all: four patient problems, four questions about each, on each of the four tools. Your score reflects how often your clinic is named, and how often your website is used as a source, across those answers.

The same question does not get the same answer twice. AI assistants do not look up a fixed list. Every time they are asked, they write a fresh answer, and the clinics they mention can change from one day to the next, even when nothing about your website or your competitors has changed. Google does not always show an AI Overview for a question either. Ask the same question an hour later and you may get a different recommendation.

So a single score is a snapshot, not a verdict. A few points up or down between two measurements is normal movement. It is not a sign that your content has stopped working.

The four patient problems are chosen to reflect your clinic. If you're a general clinic with special interests, they cover your main everyday conditions and up to two of your specialties. If your clinic focuses entirely on one area, they all come from that area. Occasionally we refresh your patient problems, for example when your services change. When that happens, your results start a new baseline. Your earlier scores stay in your history, marked as measured on the previous questions, but they aren't compared with the new ones, because they measured different things.

Why it works this way

AI assistants can only recommend what they have found and trust. A new page has to be published, found and read before it can influence their answers, and that does not happen the day you press publish.

And one article on its own is rarely enough. AI tools look for depth: a clinic that clearly knows a subject inside out. Depth comes from building out a topic properly: a main page for the topic, plus supporting pages that each answer one specific question patients ask about it. One page on knee pain is a single mention. A main knee pain page with supporting pages on running knee pain, knee pain on stairs and when to get a knee checked, all written under your named clinicians and linked together, shows real expertise in that subject.

That depth builds over time, and it builds fastest when the work follows a plan rather than being scattered across lots of unrelated topics. Each piece you publish in a topic makes the others count for more. This is how the Clinical Content System works: it works out which topics matter most for your clinic, plans the pages each topic needs in the right order, and writes them under your clinicians' names, so the depth adds up instead of being spread thin. "What are topic areas (pillars and clusters), and why does my plan build them?" explains how.

What to do now

  1. Judge progress over several measurements, not one. Look at the direction, not a single number.
  2. Look at where you are named, not just the total. Are you named on more of the questions, on more of the AI tools, and for the conditions you most want to be known for?
  3. Build depth in the topics you most want to be known for, rather than one page on each of many topics.
  4. Publish what you have written. A page that is not live on your website cannot be found by any AI tool.
  5. Make sure each clinician's name, qualifications and registration number sit on their own profile page, so patients can see exactly who is behind your advice.

The question behind the question

"Is this working?" The pages you add now are what later measurements pick up. Building a clinic's presence in AI answers is not an overnight fix, and anyone who promises a date or a guaranteed result is guessing. What makes the difference is depth, built to a plan and kept up over time. Your results show which AI tools named you, and for which questions, so when you are named more often, you can see exactly where and for what.

The next question is usually...

Still stuck? This usually means...