A Q&A with Ronn Torossian, Founder and Chairman, 5W AI Communications | Publisher, Everything-PR
Q: You just published the Health Insurance & Medicare Advantage AI Visibility Index. What is it — and why does it exist?
The Health Insurance & Medicare Advantage AI Visibility Index is the first public ranking of how 25 U.S. health insurers show up — or don't — inside the AI engines that now shape buyer decisions. We tracked 60-plus queries across five AI platforms — ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews — and measured each insurer's Citation Share: the proportion of times a brand is named when a consumer asks an AI engine a question about health insurance.
Why does it exist? Because the research surface has moved. A KFF poll found that 1 in 3 U.S. adults now use AI chatbots for health information — equal to the share using social media for the same purpose. Gallup puts it at 25% of all Americans using AI tools for health information or advice, with 59% using AI to research before doctor visits. Among adults under 30, that number is 69%. These are not early adopters anymore. This is the mainstream.
Now layer in the insurance dimension. Medicare Advantage alone covers roughly 34 million Americans and represents a $500 billion annual market. When someone asks ChatGPT "What's the best Medicare Advantage plan?" — and they are asking, millions of times — the brands that appear in the answer have a structural advantage over the brands that don't. That's what we're measuring.
Q: Walk me through the methodology. How do you actually measure Citation Share inside an AI engine?
Each query was run five times per engine in clean sessions — no login, no history, no personalization bias. We tracked 60-plus prompts across the five major AI platforms. The prompts mirror the way real consumers ask: "Best Medicare Advantage plan," "UnitedHealthcare vs. Humana," "Best health insurance for seniors," "Which Medicare plan should I pick" — and about 50 variations across plan types, comparison queries, and benefit-specific searches.
Citation Share is measured as the proportion of brand mentions across all tracked prompts. If an engine answers 100 queries and names UnitedHealthcare in 15 of them, that's a 15% Citation Share. The percentages are directional — they measure relative visibility, not absolute reach. But the directional signal is what matters. It tells you who owns the answer and who doesn't.
We also track the source architecture behind those citations — what the AI engines are drawing from. The retrieval sources fall into four buckets: health and insurance editorial (NerdWallet, Forbes Health, U.S. News, Investopedia), consumer-satisfaction data (J.D. Power, CMS Star Ratings, Medicare.gov), news coverage (Healthcare Dive, Modern Healthcare, KFF, Reuters), and brand-owned content. That sourcing layer tells you why a brand gets cited — not just whether it does.
Q: Who's winning? What does the top of the ranking look like?
UnitedHealthcare holds the #1 position with approximately 15% Citation Share. Humana is second at 12%. Kaiser Permanente — which surprises people — is third at 9%, despite a much smaller geographic footprint. Aetna (CVS Health) at 7.5%, Blue Cross Blue Shield at 6.5%, Cigna at 5%, Elevance Health (Anthem) at 4%.
Then the drop-off steepens. Centene/Wellcare at 3%. The insurtech cohort — Devoted Health (2.2%), Clover Health (1.8%), Alignment Health (1.5%) — is small in absolute terms but growing fast. Oscar, Molina, SCAN, HCSC round out the top 15, all under 1.5%.
The full index ranks all 25 insurers. Below the top 15, you see regional Blue Cross affiliates (Highmark, Florida Blue, Independence, Premera), integrated systems (UPMC), government-program specialists (CareSource, Medica), and marketplace aggregators (eHealth, GoHealth). Most of these sit below 1% Citation Share — functionally invisible inside AI answers.
Q: UnitedHealthcare lost nearly a million Medicare Advantage members heading into 2026. How does it still rank #1?
This is one of the most important findings in the study. Citation Share and enrollment share are correlated — but they don't move in lockstep. UnitedHealthcare shed approximately 930,000 MA members. That's a massive enrollment hit. But their AI citation position barely moved.
Why? Because AI engines don't cite enrollment data. They cite the editorial and informational record that accumulated over decades — news coverage, CMS filings, provider directories, consumer education content, third-party reviews. UnitedHealthcare has the largest MA footprint in the country, and that generated years of editorial, news, and reference coverage. The LLMs are drawing on that accumulated record. It's a retrieval anchor — deep, wide, and self-reinforcing.
The flip side: Humana added roughly 1.2 million MA members and saw its Citation Share climb to 12%. Growth narratives generate coverage. Coverage generates citations. That's the flywheel working in real time.
Q: You mentioned Kaiser Permanente at #3. That seems high for a regional system. What's driving that?
Consumer satisfaction data. Kaiser consistently ranks at or near the top of J.D. Power surveys and CMS Star Ratings. AI engines treat satisfaction data as a direct citation asset — when someone asks "best health insurance company," the engine pulls from the sources that rank insurers, and Kaiser dominates those sources.
Kaiser also operates an integrated model — insurance plus care delivery — which makes it a structurally distinct entity in the AI knowledge graph. It's not diluted across 30 regional affiliates the way Blue Cross Blue Shield is. Federation structures fragment Citation Share. Kaiser's unified brand concentrates it.
Q: Let's talk about the AI usage data behind this. How big is the shift toward AI-driven health research?
The numbers have moved past the tipping point. Here's what the data says:
32% of U.S. adults have used AI chatbots for health information in the past year, according to KFF — with 29% seeking physical health information and 16% seeking mental health information. That equals the share using social media for the same purpose. AI is no longer the alternative channel. It's a primary channel.
Gallup found that 25% of all Americans have used AI tools for health information or advice, with 59% using AI to research before seeing a doctor and 56% researching after visits. Among 18-to-29-year-olds, 69% are doing self-directed health research through AI. Even among adults 65 and older — the Medicare population — 43% report using AI for health research.
And the trust question is evolving. Only 4% strongly trust AI health information today, per Gallup. But 65% of users cite "quick or immediate information" as the primary reason they turn to AI, per KFF. Convenience is winning over trust — and convenience is how habits form.
The privacy dimension matters too. 41% of AI health users have uploaded personal medical records to chatbots. That's not casual browsing. That's deep engagement. And 77% of the general public expresses privacy concerns about it — which means the behavior is outrunning the comfort level. The usage is real.
Q: What are the structural dynamics reshaping citations right now? What should insurers pay attention to?
Six things are moving simultaneously in 2026:
One — the most turbulent enrollment cycle in years. Major carriers entering and exiting counties. UnitedHealthcare losing nearly a million members. Humana gaining 1.2 million. Devoted Health more than doubling. Alignment Health growing 21%. Every enrollment shift generates a wave of news coverage — and that coverage reshapes what AI engines cite.
Two — county exits are reshaping local AI answers. When a carrier pulls out of a county, the "best plan in [county]" query changes overnight. The carrier that fills the gap inherits the citation.
Three — benefit cuts entering comparison queries. CMS benefit adjustments for 2026 are generating editorial coverage, and that coverage shapes how AI engines frame head-to-head comparisons.
Four — Star Ratings are a direct citation driver. CMS Star Ratings and J.D. Power surveys are among the most-cited sources in AI answers about insurance quality. A Star Rating change moves Citation Share.
Five — seniors are initiating research via chatbots, not brokers. The traditional Medicare enrollment path ran through brokers and direct mail. AI engines are disintermediating that channel. Broker and mailer channels remain invisible to AI — they generate zero citations.
Six — insurtechs are claiming "newer option" query ownership. Devoted Health, Clover, Alignment, Oscar — they're small in enrollment but disproportionately visible in "new" and "alternative" insurance queries because their growth narratives generate editorial coverage.
Q: What should a health insurer do about this? What's the playbook?
Ten things. All operational, not theoretical:
Audit your AI Citation Share quarterly — weekly during open enrollment. You cannot manage what you don't measure. Most insurers have no idea how they appear inside ChatGPT, Claude, or Perplexity. Start there.
Treat CMS Star Ratings and J.D. Power as citation assets, not compliance checkboxes. These are among the top sources AI engines pull from. A Star Rating improvement is a GEO event.
Build content targeting high-intent enrollment queries. "Best Medicare Advantage plan in [state]." "UnitedHealthcare vs. Humana." "Is Kaiser Permanente good?" These are the prompts consumers are typing into AI engines. Your content should be structured to answer them — tables over prose, direct answers in the first 30% of the page, named-expert attribution.
Recognize that broker and mailer channels don't move citations. They move enrollment. They don't move AI visibility. Both matter, but if your entire acquisition strategy is broker-driven, you're invisible to the fastest-growing research channel.
Consolidate brand architecture. Federation structures — looking at you, Blue Cross Blue Shield — dilute Citation Share across dozens of entities. AI engines can't aggregate what they can't identify as the same brand.
Treat rebrands as citation events. Wellpoint's rebrand reset its citation history. Without a deliberate content push, a rebrand is a citation wipeout.
Publish transparent benefit and cost content. AI engines favor content that directly answers cost-comparison queries. Opaque pricing pages don't get cited. Structured, comparable data does.
Build location-specific content following county exits. When a competitor exits a county, publish content that answers the "who replaced them" query. That's a citation opening.
Earn the health-editorial record systematically. NerdWallet, Forbes Health, U.S. News, Investopedia — these are the editorial sources AI engines trust most for insurance queries. Earned coverage in these outlets is a direct Citation Share driver.
Re-audit after every Star Ratings release and enrollment cycle. Citation Share is not static. It shifts with every news cycle. The insurers that monitor and respond in real time will own the answer during open enrollment — the highest-intent search window in American healthcare.
Q: You've said Citation Share is "the new market share." That's a big claim. What makes you confident?
Because the research behavior has already shifted and the data confirms it. One in three American adults is using AI for health information right now. Not next year. Now. Gallup says 14% of AI health users — roughly 14 million Americans — have skipped a provider visit entirely based on what an AI chatbot told them.
When a consumer asks ChatGPT "Which Medicare Advantage plan is best for me?" — the brand that appears in the answer has a structural advantage over every brand that doesn't. That's not an impression. It's not a click. It's a recommendation from a trusted intermediary at the moment of decision. That is market share. The insurers who understand this will build for it. The ones who don't will wonder why their enrollment numbers moved and their broker can't explain it.
Q: Where can people access the full study?
The full Health Insurance & Medicare Advantage AI Visibility Index is published at 5wpr.com/ai-visibility-index/health-insurance-ai-visibility-index-2026. The broader 5W AI Visibility Index — covering multiple industries — is at 5wpr.com/ai-visibility-index. Both are open access.
