In June, MGMA asked 218 medical practices a question almost nobody had been asking: do you have a strategy for AI search visibility? Put plainly — when a patient asks ChatGPT, Google AI or Perplexity which local practice treats a condition or performs a procedure, are you part of the answer?
Fifty-five percent said no. Thirty-five percent said yes. Eleven percent were unsure.
of medical practices reported having no strategy for AI search visibility. The barriers respondents named were time, money, staffing, IT support — and ownership.
Source: MGMA Stat poll, 218 applicable responses, 16 June 2026
That 55% is the number being quoted. It is the wrong number to focus on.
The finding inside the finding
MGMA looked at what the 35% who claimed a strategy actually described. Many of them were describing general AI adoption — using AI tools in the practice — rather than anything to do with search visibility. The responses that were genuinely on-topic clustered around updated website content, structured physician bios, blog posts, Google Business Profile and social updates, patient reviews, and support from a marketing vendor.
Which means the real share of practices with a deliberate AI search visibility strategy is meaningfully lower than 45%. Some portion of the confident third were answering a different question.
The gap is wider than the headline says, and the people inside it do not know they are in it.
That matters more than a percentage point, because it changes what kind of problem this is. A market where 55% know they have a gap is a market with demand. A market where a further slice believes it has solved something it has not yet started is a market with a definition problem — and definition problems are much earlier, and much more open, than demand problems.
Ownership is the barrier that explains the rest
The list of stated obstacles is worth reading closely: time, money, staffing, IT support, and ownership. The first four are the barriers every initiative faces. The fifth is the one that actually explains the data.
In a typical practice or health system, nobody's job description contains this. It is not clearly marketing, because it is not a campaign and there is no channel to buy. It is not clearly IT, because the fix is mostly content and data consistency rather than infrastructure. It is not clearly operations, though it affects patient acquisition. It sits in the gap between functions, which is precisely where work goes to not happen.
This is also why the practices that have moved tend to have done so through a vendor rather than internally. Not because the work is technically hard, but because assigning it internally requires someone to decide it belongs to them.
What a strategy would actually consist of
The responses MGMA collected are not wrong — updated bios, Google Business Profile, reviews and structured content all matter. They are just partial, and they are the parts that look most like traditional marketing. The fuller picture in healthcare has more layers, because generative systems draw far more heavily on the directory and registry substrate than they do on any practice's own website.
A complete position covers eight dimensions: clinical authority, institutional authority, service line authority, physician expertise, third-party validation, content authority, entity clarity, and the measured visibility outcome itself. The first seven are inputs. The eighth is the result — which means a weak position is always traceable to specific dimensions, and those dimensions can be assigned to specific people.
Entity clarity is usually the largest and most fixable of them. If a practice's name, address, specialty taxonomy, affiliations and service listings disagree across NPPES, Google Business Profile, Healthgrades, Vitals and the payer directories, a generative system resolving that ambiguity will frequently resolve it in favour of a competitor whose record is clean. That is not a marketing failure. It is a data hygiene failure with a marketing consequence, and it is invisible until somebody measures it.
The one thing worth doing before anything else
Measure where you actually stand. Not a screenshot of one ChatGPT answer — a proper measurement, run repeatedly, across the platforms your patients actually use, scored against the competitors you actually have.
Most practices in that 55% assume they are invisible. Some are not. Some are visible for one service line and absent for another. Some appear regularly but are described inaccurately, which is a different and more urgent problem. You cannot know which of those you are without measuring, and every one of them calls for a different response.
Related
Find out where your practice actually stands
A Share of Discovery assessment measures where you appear across ChatGPT, Google AI, Gemini and Perplexity — and which competitors are being recommended instead.