MedTech AI Visibility: What AI Engines Actually Cite When Buyers Ask About Medical Device Companies

Medical device companies spend years on FDA clearance and millions on clinical evidence, and a hospital procurement team's first read of the category now comes from an AI assistant. A common argument, drawn from case studies, holds that the assistant reads Forbes, STAT News and MedTech Dive rather than your 510(k) summary or your ClinicalTrials.gov record. The Machine Relations Index can now measure that argument, and the measurement is sharper than the argument. In the September 19, 2026 release, across 15,883 monitored answer runs, fda.gov was cited in 7 and clinicaltrials.gov in 2. The regulatory record is not what the engines read. But neither, on the buyer questions, is the press: the three healthcare buyer questions the Index has published are led on every one by a healthcare software company's own domain, and the outlets a MedTech PR program is built to earn hold, between all of them, one slot inside a buyer top 15. The source the engines do cite for healthcare evidence is the National Institutes of Health, at #1 on healthcare news questions and inside the top 15 on how buyers choose. This page covers what that means for a medical device company: which surfaces the engines read, where an earned placement lands, and the two-track plan that fits.

Key Takeaways for MedTech AI Visibility

• The Index has no medical-device category yet. The nearest measured category is healthcare services, with four question shapes published and three collecting. Its pattern is the one a device company should plan against until a device segment publishes: buyer questions are answered from vendors' own pages, news questions from NIH and the healthcare trade press.
• The regulatory record is barely a source. fda.gov is cited in 7 of 15,883 runs across the whole Index (6 of them healthcare news, #175 of 1,264) and clinicaltrials.gov in 2, in no healthcare segment. NIH, which hosts peer-reviewed evidence at PubMed Central, is the most-cited healthcare news source at 25.12% of 613 runs.
• Of the seventeen outlets this page named as the MedTech publication ecosystem, Becker's Hospital Review (#9 of 199 on best tools), Forbes (#16) and Healthcare IT News (#90) are the only three inside a published healthcare buyer top 100, and Becker's is the only one inside a top 15. MedTech Dive is #70 of 1,264 on healthcare news, STAT News #71, Reuters #1,106, TIME #1,180. Fortune and USA Today are not observed in any healthcare segment.
• The three PR case studies this page cited (Stereotaxis, an RH Strategic device launch, Cognixion) still say what they said: 110-plus pieces, 860 placements, 35-plus placements. What the Index adds is where those placements land in AI answers, and for the device makers themselves the answer is that stereotaxis.com and cognixion.com are outside the 22,213-domain release.
• The plan is two tracks: an owned evidence page per buyer question, written to the compliance line, because that is what the buyer answers are made of; and the healthcare trade press plus NIH-indexed publication for the news answers, where the press is actually cited.

What the Regulatory Record Reaches in AI Answers

The claim this page has always made is that an AI engine does not read your 510(k) submission or your ClinicalTrials.gov registration. The Index monitors six answer engines (ChatGPT, Claude, Gemini, Google AI Mode, Google AI Overviews and Perplexity) on buyer prompts and records which root domains each answer cites, and release mri_score_v2.0+2026-09-19+0cad03121f60 covers 2026-05-10 to 2026-09-19, 15,883 monitored runs and 22,213 cited domains. The government and academic sources a device company's evidence lives on:

Source	Cited runs, whole Index (of 15,883)	Healthcare services standing
nih.gov	Most-cited healthcare news source	News-driven citations #1 of 1,264 (154 of 613, 25.12%); how buyers choose #14 of 214 (8 of 107, 7.48%); leads the collecting is-it-worth-it shape at 28 of 101 runs
cms.gov	Observed on two healthcare shapes	How buyers choose #36 of 214 (4 of 107); news #20 of 1,264 (25 of 613, 4.08%)
hhs.gov	Observed on healthcare news	News #31 of 1,264 (19 of 613, 3.10%)
fda.gov	7 (0.04%), five engines, collecting	News #175 of 1,264 (6 of 613, 0.98%); not observed on any buyer shape
clinicaltrials.gov	2 (0.01%), two engines, collecting	Not observed in any healthcare segment

Read that as a device company. The FDA's 510(k) database and ClinicalTrials.gov are the two most authoritative public records of your product, and between them they are cited in nine answer runs out of 15,883. The peer-reviewed literature is a different matter: nih.gov, which serves PubMed Central, is the single most-cited source on healthcare news questions and the only academic or government source inside a published healthcare buyer top 15. Clinical evidence gets cited when it is published where the engines already read, and the engines read the journal record at NIH far more than they read the regulator.

Citation quality in medicine is a measured problem in its own right: the MedCite framework, built by researchers at the University of Illinois, NIH, the University of Virginia and Microsoft, was designed to evaluate how language models generate and verify citations for medical tasks (Wang et al., ACL 2025). The Index measures the other side of the same problem: not whether the citation is correct, but which domains get cited at all.

Where the MedTech Publication Ecosystem Lands

Medical device PR plans usually list the outlets that matter in three tiers: mainstream authority (Forbes, Business Insider, Reuters, TIME, Fast Company, USA Today), business and technology (Inc., Fortune, Entrepreneur, TechCrunch, AP News), and healthcare trade (STAT News, MedCity News, Fierce Healthcare, Healthcare IT News, MedTech Dive, Becker's Hospital Review). Each outlet has its own Index profile, which lists every category and shape the domain was observed in. Here is where each lands on healthcare questions:

Outlet	Cited runs, whole Index (of 15,883)	Healthcare buyer-question standing (published shapes)	Healthcare news-driven standing (of 1,264)
Forbes	657 (4.14%), Confidence A, six engines	Best tools #16 of 199 (7 of 107, 6.54%); top lists #164 of 223 (1 of 102); not observed on how buyers choose	#47 (15 of 613, 2.45%)
Business Insider	99 (0.62%), Confidence C, five engines	Not observed on a published buyer shape (1 problem-first observation, collecting)	#107 (9 of 613, 1.47%)
Reuters	13 (0.08%), three engines, collecting	Not observed	#1,106 (1 of 613)
TIME	56 (0.35%), Confidence C, four engines	Not observed	#1,180 (1 of 613)
Fast Company	29 (0.18%), three engines, collecting	Not observed	#836 (1 of 613)
USA Today	19 (0.12%), three engines, collecting	Not observed	Not observed
Inc.	11 (0.07%), four engines, collecting	Not observed	#494 (2 of 613)
Fortune	59 (0.37%), Confidence C, four engines	Not observed	Not observed
Entrepreneur	18 (0.11%), four engines, collecting	Not observed	#814 (1 of 613)
TechCrunch	164 (1.03%), Confidence B, five engines	Not observed	#576 (2 of 613)
AP News	7 (0.04%), one engine, collecting	Not observed	#662 (1 of 613)
STAT News	23 (0.14%), four engines, collecting	Not observed on a published shape; 11 of 101 on the collecting is-it-worth-it shape	#71 (12 of 613, 1.96%)
MedCity News	33 (0.21%), Confidence C, six engines	Not observed	#14 (33 of 613, 5.38%)
Fierce Healthcare	123 (0.77%), Confidence B, six engines	Not observed on a published shape (3 is-it-worth-it and 2 problem-first observations, collecting)	#2 (117 of 613, 19.09%)
Healthcare IT News	9 (0.06%), four engines, collecting	Best tools #90 of 199 (2 of 107)	#148 (7 of 613, 1.14%)
MedTech Dive	12 (0.08%), two engines (Claude, Google AI Mode), collecting	Not observed	#70 (12 of 613, 1.96%)
Becker's Hospital Review	50 (0.31%), Confidence C, six engines	Best tools #9 of 199 (13 of 107, 12.15%)	#13 (35 of 613, 5.71%)

Four device-specific outlets this page did not name, read the same way: Modern Healthcare is #39 on healthcare news (17 of 613, 2.77%) and all 17 of its whole-Index citations sit in that one segment; MassDevice is #268 (4 of 613); Fierce Biotech is #176 (6 of 613); Medical Device Network is outside the 22,213-domain release. CBS News, which appears in the Cognixion case below, was cited in 45 runs across the whole Index and reaches healthcare only on the collecting comparisons shape (5 of 96).

Read together: the trade layer this page called uniquely important is real on news-shaped questions, where Fierce Healthcare is #2, Becker's #13, MedCity #14 and Modern Healthcare #39, and the trade title with "MedTech" in its name is cited on Claude and Google AI Mode only, twelve times in the whole window. On the buyer questions, Becker's at #9 on best tools and Forbes at #16 are the whole editorial presence of this list. Reuters and TIME, the two names an FDA-approval release is written to reach, are cited once each on healthcare news across 613 runs.

"Not observed" means the Index recorded no citation of that domain in that segment during the window; it is a statement about these monitored prompts on these six engines, not about every AI answer. Confidence tiers are domain-level; segment rows carry no tier of their own, and a citation is not evidence that the answer was supported by the source.

What Healthcare Buyer Answers Are Made Of

The engines are citing something on healthcare buyer questions, and it is neither the regulator nor the press. The healthcare services category has four shapes published in this release and three collecting, each with its own denominator:

Question shape	State	Runs	Most-cited source	Rate	What the top ten looks like
Best tools	Published	107 on 7 dates	omnimd.com	32.71%	CureMD #2, MDClarity #3, OneMed Billing #4, PracticeSuite #5, Certify Health #6; Becker's #9 is the one editorial row
How buyers choose	Published	107 on 7 dates	omnimd.com	14.95%	Elevate PFS #2, Phreesia #3, Centauri #4, NexHealth #5; NIH #14 is the first non-vendor row
Top lists	Published	102 on 7 dates	omnimd.com	24.51%	Commure #2, Medical Billers and Coders #3, EHR in Practice #4, Guideflow #6, Praxis EMR #7; HIPAA Journal #11
News-driven citations	Published	613 on 53 dates	NIH	25.12%	Fierce Healthcare #2, Galen Growth #3, AHA #4, LinkedIn #5, Y Combinator #6, Healthcare Dive #7, Rock Health #8, athenahealth #10
Is it worth it	Collecting	101 on 6 dates	NIH, 28 observations	No rate published	AMA 14, Freed 13, Commure 12, STAT News 11
Problem-first research	Collecting	102 on 6 dates	Medical Billers and Coders and YouTube, 10 each	No rate published	Vendor pages and community platforms
Comparisons	Collecting	96 on 5 dates	adsc.com, 16 observations	No rate published	Vendor pages

On the three published buyer shapes, the Index's unclassified "other observed source" bucket, which is where healthcare software companies cited on their own domains land, holds 74, 87 and 78 of the three top 100s; editorial publications hold 8, 1 and 7. The same domain, omnimd.com, an EHR and practice-management company's site, leads all three. These are healthcare software and services prompts, not surgical-robotics prompts, so the top tens are practice-management, billing and patient-intake companies rather than device makers, and the device makers this page has used as examples confirm the segment's scope: medtronic.com was cited twice across the whole Index (healthcare news #522), and stereotaxis.com, intuitive.com and stryker.com are outside the 22,213-domain release. A device buyer's question ("best Class II surgical robots") is not yet a monitored shape. What the healthcare category establishes is the pattern that shape will almost certainly follow, because it is the pattern every published category follows: the buyer answer is assembled from pages that answer the buyer's question in plain terms, and those pages are overwhelmingly the vendors' own. The full seven-shape reading is on the healthcare hub.

What a MedTech PR Program Earns, and Where It Lands

Three case studies have anchored this page, and each was re-read at its source on September 19, 2026. They still say what they said. The Index now says where it lands.

Stereotaxis and MWW Health. MWW's case study reports that the surgical robotics company nearly doubled its earned media in 2024, securing 110-plus pieces of unique earned editorial coverage, expanded its media mix from 14 publications in 2024 to more than 35 outlets in 2025 including Modern Healthcare, MedTech Dive and Forbes, and established its CEO and key opinion leaders as authority voices with over 35 pieces of thought-leadership coverage. Where those three outlets land on healthcare questions: Forbes #47 on news and #16 on best tools, Modern Healthcare #39 on news, MedTech Dive #70 on news. Stereotaxis's own domain is outside the release.

A breakthrough device launch and RH Strategic. RH Strategic's case study reports 860 media placements over two years, coverage in Forbes, Reuters and TIME, and seven award wins, for a first-of-its-kind breakthrough medical device nearing FDA approval, with the program running from pre-launch positioning through FDA clearance and commercial rollout. The detail that still matters is "over two years": the narrative was set before the approval arrived. Where the named outlets land: Forbes #47 on healthcare news, Reuters #1,106, TIME #1,180.

Cognixion and Escalate PR. Escalate's case study reports 35-plus earned media placements in six months for the brain-computer interface company, across outlets including Forbes (exclusive), CBS News 24/7 (exclusive), MedTech Spectrum, Medical Device Network and HIMSS TV, built on the company's non-invasive approach, clinical research and FDA breakthrough designation. Where those land: Forbes as above, CBS News on the collecting healthcare comparisons shape only, Medical Device Network outside the release. Cognixion's own domain is outside the release.

Two readings, both true. First, these programs did what PR programs do: they produced coverage a human buyer, an investor and an analyst read, and the Stereotaxis study reports a stock re-rating alongside the coverage. Second, none of the outlets those programs earned is a meaningful source on the buyer questions the Index has published in healthcare, and the companies' own domains are not in the release at all. A placement is a human surface and, on news-shaped questions in the trade press, an engine surface. It is not, on this evidence, a buyer-answer surface.

Why Milestone PR Fails Medical Device Companies

Traditional MedTech PR runs on a milestone model: FDA clearance, funding round, product launch, repeat. Between milestones, silence. That model worked when buyers did their own research through sales reps, trade shows and analyst reports. It fails when an assistant is summarizing your category every day from a small set of sources it already cites for that shape of question, and nothing of yours is in that set.

The Index shows what "in that set" means in healthcare. News-driven answers are assembled from NIH, Fierce Healthcare, Galen Growth, the AHA, LinkedIn, Y Combinator, Healthcare Dive and Rock Health. Buyer answers are assembled from healthcare companies' own pages. A device company with a press-release archive and a HIMSS booth has, on the engines' evidence, neither a presence in the trade press that leads the news answers nor an owned page that states a shortlist, a comparison or a cost, and the milestone release itself is a wire document: Machine Relations' concentration study on the same Index found the whole wire and press-release layer is nine domains cited 226 times in 125 days, 0.20% of every citation recorded.

The difference between the companies that appear and the ones that do not is not budget. It is which surfaces the program builds.

The Compliance Constraint, Read Against the Evidence

Health tech PR sits at the intersection of three regulated audiences, patients, providers and payers, with HIPAA, FDA marketing guidance and FTC substantiation standards governing what can be said (Everything PR). Most device companies see those constraints and freeze into generic language that says nothing.

The Index gives a reason to do the opposite. The sources the engines cite most on healthcare questions are specific, sourced and verifiable: NIH-hosted research at 25.12% of news runs, CMS and HHS inside the news top 35, the AMA and STAT News leading the collecting is-it-worth-it observations. Content framed around cleared indications, peer-reviewed data and named clinical evidence is the kind of content those sources are made of. 4media Group's FDA approval communications roadmap describes building the pre-approval foundation on audience research and segmentation rather than premature claims; the Index adds that the foundation should include an owned page per buyer question, written to the compliance line, because that page is what the buyer answer will be assembled from.

A Two-Track Plan for a Medical Device Company

Track one: an owned evidence page per buyer question. The healthcare buyer answers are made of vendors' own pages, so stand up one page per shape you need to win, each of which states its answer in plain, compliant terms: a shortlist page for best tools that names the cleared indications and the evidence behind each criterion; a comparison page for the head-to-head your buyers actually run; a cost, reimbursement and payback page for is it worth it, the shape NIH and the AMA currently lead; and a problem-first page in the clinical vocabulary the buyer uses before they know your product name. Publish the evidence where the engines read it: the peer-reviewed record indexed at NIH is cited 154 times on healthcare news and 8 times on how buyers choose, while the 510(k) database is cited 7 times across the entire Index. Check your robots.txt before any of this: Google's crawler documentation says the Google-Extended token governs Gemini grounding rather than Search inclusion (Google: common crawlers), and a page an engine cannot fetch is a page it cannot cite.

Track two: the trade press for the news answers, and the mainstream press for the humans. The news-shaped healthcare answers cite the trade press in order: Fierce Healthcare #2 (117 of 613), Becker's #13, MedCity News #14, Modern Healthcare #39, MedTech Dive #70, STAT News #71, Fierce Biotech #176, MassDevice #268. That order is your pitch order for the citation, and it is not the order a MedTech agency's tier list runs. Forbes, Reuters, TIME, CBS News and TechCrunch remain worth earning for the reader who is a person: the hospital CFO, the investor, the acquirer. Earn them for that reader, and point every placement at the owned page that answers the buyer's question, so the human and the engine both have somewhere to land.

Entity consistency across both tracks. If your website says "AI-assisted surgical guidance," your release says "robotic-assisted surgery platform" and your executives' profiles say "intelligent surgical robotics," an engine building an entity from your footprint has three stories. Pick one, use it on the owned pages, in the trade coverage and in the journal record, and keep it.

Common Mistakes MedTech Companies Make

Waiting for FDA clearance to start. Both device-launch case studies above ran pre-launch. By the time you have clearance, the sources the engines cite for your category are already set.

Treating the tier list as the citation list. The outlets ranked highest for prestige (Reuters, TIME, USA Today, Fortune) are cited once, once, never and never on healthcare news in 613 runs. The outlets cited most (Fierce Healthcare, Becker's, MedCity) were the "trade" tier.

Publishing clinical evidence only in the regulatory record. The 510(k) summary and the trial registration are cited nine times across 15,883 runs between them. Peer-reviewed publication indexed at NIH is the evidence surface the engines read.

Having no owned page for the buyer's question. Coverage describes you; it rarely answers "which is the best" or "is it worth it." In healthcare, every published buyer shape is led by a company answering that question on its own domain.

Confusing volume with authority. Fifty press releases a year put fifty documents on a layer the Index measures at 0.20% of all citation, or 0.37% once PR Newswire is counted as wire distribution rather than the editorial publication the release files it as. Ten pieces of evidence-backed editorial in the outlets the engines cite, plus four owned pages, is a different program.

What a MedTech AI Visibility Program Looks Like

Traditional MedTech PR	Machine Relations for MedTech
Milestone press releases	Owned evidence page per buyer question, written to the compliance line
Trade show appearances	Trade press placements in the order the Index cites them
Product feature announcements	Peer-reviewed evidence published where the engines read it
One-off mainstream placements	Mainstream placements for the human reader, each pointing at an owned page
Generic category language	One entity story across site, coverage and journal record
Silence between milestones	A release read each week to see which shape moved
Sales-led follow-up	Pre-qualified buyers arriving from an answer that named you

A hospital procurement officer in 2026 starts with a query, not a sales call. If your company is not in that answer, you are not on the shortlist. Clinical evidence alone does not fix this, because the record it lives in is barely read. A strong press program alone does not fix this, because the outlets it earns are not the buyer-answer sources. The two together, pointed at owned pages that state the answer, is what Machine Relations does for a device company.

What Machine Relations Means for Medical Device Companies

Machine Relations is the discipline of earning AI citations and recommendations for a brand by making that brand legible, retrievable, and credible inside AI-driven discovery; it is the parent category for GEO, AEO, AI SEO and AI PR. For a device company it means treating the editorial and evidence footprint as infrastructure the engines read, not a collection of clippings: a clinical credibility layer anchored in cleared indications, peer-reviewed data and named experts; a structured entity chain, one name and one category across every surface; and a continuous presence in the sources the release shows are cited, measured against that release rather than against a tier list.

The Index is published by Machine Relations and is not an AuthorityTech product. It measures observed root-domain citations in monitored prompts; it does not measure recommendation quality, traffic or revenue, and correlation between a placement and a later citation stays correlation until a release shows otherwise.

FAQ

Can medical device companies get featured in Forbes or TechCrunch?

Yes, and the case studies on this page show programs that did. Know what the placement reaches: Forbes is #47 of 1,264 on healthcare news and #16 of 199 on best tools, the strongest general-interest position in the category; TechCrunch is #576 on healthcare news and on no healthcare buyer shape. Earn them for the human reader and the investor.

Does FDA clearance or a ClinicalTrials.gov listing make an AI engine cite you?

Not on this evidence. fda.gov is cited in 7 of 15,883 monitored runs across the whole Index and clinicaltrials.gov in 2, and neither appears on a published healthcare buyer shape. The evidence surface the engines read is the journal record at NIH, #1 on healthcare news questions.

How does FDA approval timing affect earned media strategy?

Start pre-approval. Both device-launch case studies on this page ran from pre-launch positioning through clearance and rollout, and the Index adds the reason: the sources an engine cites for a category are set by the time your announcement lands, so the owned pages and the trade coverage need to exist first.

Which publications should a medical device company target for AI visibility?

For healthcare news answers, the trade press in the order the Index cites it: Fierce Healthcare (#2 of 1,264), Becker's (#13), MedCity News (#14), Modern Healthcare (#39), MedTech Dive (#70), STAT News (#71). For buyer questions, there is no publication to target: every published healthcare buyer shape is led by companies cited on their own domains, with Becker's #9 and Forbes #16 on best tools as the editorial exceptions.

How do AI engines decide which medical device companies to cite?

The Index measures which domains are cited, not why. What it shows in healthcare is that buyer answers are assembled from pages that state the answer in plain terms, and those pages are the companies' own; news answers are assembled from NIH and the healthcare trade press. A device company that has no page stating its answer and no presence in those sources is not in the answer, whatever the product.

Is press release distribution enough for MedTech visibility?

No. The wire and press-release layer is nine domains cited 226 times in 125 days across the whole Index, 0.20% of all citation as the release classifies its members, or 0.37% once PR Newswire is counted as wire distribution rather than the editorial publication the release files it as. A release is the record and the discipline; the placement and the owned page are the surfaces.

Is there a medical-device segment in the Machine Relations Index?

Not yet. The healthcare services category is the nearest measured category, and this page is written against it. When a device-specific shape publishes, this page will carry its ranks.

Sources and method

Every Index figure on this page is read from the live release pages at machinerelations.ai on September 19, 2026, release mri_score_v2.0+2026-09-19+0cad03121f60 (window 2026-05-10 to 2026-09-19, 15,883 monitored runs, 22,213 cited domains): the healthcare services category page and its seven segment pages (best tools, how buyers choose, top lists, news-driven citations, is it worth it, problem-first research, comparisons), and the domain profiles of nih.gov, cms.gov, hhs.gov, fda.gov, clinicaltrials.gov, forbes.com, businessinsider.com, reuters.com, time.com, fastcompany.com, usatoday.com, inc.com, fortune.com, entrepreneur.com, techcrunch.com, apnews.com, cbsnews.com, statnews.com, medcitynews.com, fiercehealthcare.com, healthcareitnews.com, medtechdive.com, beckershospitalreview.com, modernhealthcare.com, massdevice.com, fiercebiotech.com and medtronic.com, each at https://machinerelations.ai/index/domains/<domain>. Segment standings are the release's own rank of total for each domain and segment; "not observed" is the absence of that domain from that segment's observed citations in the window, enumerated per domain at its profile; "outside the release" means the profile URL returns 404 for that domain (medicaldevice-network.com, stereotaxis.com, cognixion.com, intuitive.com, stryker.com, each fetched September 19, 2026). The wire-class total is from Machine Relations' citation concentration study. The three agency case studies are quoted from their own published pages, linked inline and re-read September 19, 2026; they are the agencies' accounts of their own work.