How AI Search Is Recommending Online Doctors

Patient discovery of telehealth services is shifting from search engine result pages to AI-generated answers. When a patient asks an AI assistant which online doctor service to use, the response is no longer a neutral list of options. It is a curated shortlist that shapes the entire consideration set, and brands that appear in that shortlist gain disproportionate access to patient demand at the moment intent is highest.

The August 2026 LLM Authority Index benchmark for online doctors reveals a market where raw visibility no longer equals commercial advantage. Across 560 eligible observations drawn from six AI platforms, the analysis found that recommendation power is concentrating among a small group of brands while several established names fail to convert presence into shortlist placement. CiteWorks Studio is interpreting this benchmark to show where AI recommendations are formed, which brands are winning the shortlist, and what the evidence suggests about the source patterns behind those outcomes.

Methodology

1. Market studied: Online doctors and telehealth services, including virtual consultations, online prescriptions, and digital healthcare platforms.
2. Brands/entities included: Amwell, Doctor on Demand, HealthTap, K Health, Lemonaid Health, LiveHealth Online, MDLive, Sesame, Teladoc, and PlushCare. The universe covers ten named brands and may not include all market participants.
3. Data collection date/window: August 2026, with extraction dated August 1, 2026.
4. AI platforms tested: ChatGPT, Copilot, Gemini, Google AI Mode, Google AI Overviews, and Perplexity.
5. Number of prompts tested: 800 total prompts were evaluated. 560 eligible prompts generated the observations analyzed in this report. Prompt count per individual platform was not provided in the public dataset; observations were analyzed as the primary unit.
6. Prompt categories: The public benchmark covers the consideration-stage cluster labeled "Best Online Doctors and Top Telehealth Services." The full LLM Authority Index report includes evaluation-stage comparison prompts and decision-stage pricing prompts. Findings in this report reflect the consideration-stage cluster unless otherwise noted.
7. Definition of a mention: A mention means the company appeared in an AI-generated response, regardless of framing, position, or recommendation quality.
8. Definition of a valid recommendation: A valid recommendation is a positive, shortlist-quality recommendation or ranked recommendation that earns recommendation credit in the benchmark scoring. This is the key CiteWorks distinction: visibility is not the same as recommendation credit. A brand can be mentioned and still receive no valid recommendation credit.
9. Ranking/scoring metrics used: Valid recommendation coverage, top-three rate, rank-one rate, top-ten rate, average recommended rank, net sentiment score, and positive visibility rate. Monetary metrics from the source dataset are omitted from this public version.
10. Limitations: This is a point-in-time benchmark. AI outputs can change based on platform updates, source material changes, and user context. The public version of this report omits monetary metrics present in the full LLM Authority Index dataset. This report is not a full audit or a full market census. Prompt count by platform was not provided in the supplied data.

Key Findings

Recommendation power is concentrated among three brands, with the remaining seven trailing by wide margins. Doctor on Demand, Sesame, and Teladoc account for the majority of valid recommendation credit in the consideration-stage cluster. Sesame leads valid recommendation coverage at 45.2%, followed by Teladoc at 41.8% and Doctor on Demand at 41.1%. The seven remaining brands range from 26.3% down to near-zero coverage, suggesting that AI systems are consistently advancing the same small set of providers regardless of which platform is generating the response.

Teladoc leads rank-one placement by a wide margin but does not convert that strength into category-leading recommendation coverage. The benchmark shows Teladoc holds the strongest rank-one position in the category at an 18% rank-one rate with 101 rank-one placements, more than double any competitor. Its average recommended rank of 1.92 is the best in the market. However, its valid recommendation coverage of 41.8% places it third, indicating a meaningful gap between winning the top position when recommended and being recommended consistently enough across all observations.

Amwell represents the category's most significant visibility-to-recommendation gap. Amwell appears in 35.4% of AI responses, placing it fourth in raw presence among the ten brands measured. Yet its valid recommendation coverage is 20.2%, a conversion gap of more than 15 percentage points. Its rank-one rate of 0.5% means it is almost never the first recommendation, and its top-three rate of 8.2% means it rarely appears in the critical first three positions. The dataset marks Amwell as visible but commercially weak in recommendation-stage contexts.

Sesame holds the strongest framing quality in the category. Sesame's net sentiment score of 0.82 is the highest among the major players, indicating that when the brand appears in an AI response, it is framed positively. Combined with 253 valid recommendations and a 26.4% top-three rate, Sesame is the most consistent recommendation-weighted winner in the benchmark. The combination of recommendation volume, top-three placement, and positive framing is commercially significant in a trust-dependent category like healthcare.

Platform performance varies substantially for the category leaders, suggesting uneven source coverage across AI systems. Doctor on Demand reaches 55.1% recommendation coverage in Google AI Mode and 53.9% in Google AI Overviews. Sesame reaches 50.6% in Google AI Mode and 57.6% in Google AI Overviews. Teladoc's strongest platform performance is 45% in Google AI Mode. These platform differences suggest that source material is being weighted differently across AI systems and that no single brand dominates uniformly across all six platforms.

What Changed in the Market

Patients are no longer moving only from Google results to brand websites. They are asking AI systems to compare telehealth providers, explain which services accept insurance, summarize pricing, surface alternatives, and recommend shortlists. The discovery moment has shifted from browsing to asking, and the AI-generated response now functions as the first filter in the patient journey. That filter is not neutral. It ranks, frames, and endorses, and the brands that appear at the top of those AI-generated shortlists gain a structural head start in the consideration process.

For a trust-dependent category like online healthcare, this shift carries particular weight. Patients are making decisions that directly affect their medical care, and they expect AI systems to surface providers that are legitimate, credible, licensed, and well-reviewed. The benchmark shows that AI systems are weighing third-party validation, review signals, and regulatory trust markers when deciding which brands to advance. Providers that have built strong, consistent, and credible public evidence layers are being rewarded with recommendation credit. Providers that are recognized as market participants but lack that evidence depth are being mentioned without being advanced.

The commercial consequence is direct. Brands can lose recommendation-stage visibility even when they are visible in AI answers. A brand can appear in an AI response as a factual reference, a comparison anchor, or a category example, but that does not mean the AI is recommending it. The distinction between being mentioned and being advanced is now the defining competitive metric in this category.

The market structure that the benchmark reveals is not temporary. As more patients use AI assistants for health-related discovery, the brands that consistently appear in valid recommendations will accumulate more review content, more comparison coverage, and more editorial authority. The concentration effect is self-reinforcing, and the gap between recommendation leaders and the rest of the market is likely to widen over time if no remediation action is taken.

What the Benchmark Found

Raw visibility leaders. Teladoc leads the category in raw mention presence at 67.3%, followed by Doctor on Demand at 63.6% and Sesame at 57.1%. These three brands are the most frequently named in AI responses. However, presence alone does not determine commercial outcome and should not be interpreted as recommendation strength.

Valid recommendation leaders. Sesame leads in valid recommendation coverage at 45.2%, followed by Teladoc at 41.8% and Doctor on Demand at 41.1%. Sesame also holds the highest valid recommendation count at 253, with Teladoc at 234 and Doctor on Demand at 230. The gap between these three and the rest of the field is substantial.

Top-three leaders. Teladoc leads the top-three rate at 29.5%, followed by Sesame at 26.4% and Doctor on Demand at 22.5%. These three brands dominate the critical first three positions in AI-generated shortlists. The top-three position is commercially significant because it corresponds to the brands a patient is most likely to actively consider after receiving an AI recommendation.

Rank-one leader. Teladoc is the clear rank-one leader with an 18% rank-one rate and 101 rank-one placements. Sesame follows at 9.8% with 55 placements, and Doctor on Demand at 6.8% with 38 placements. No other brand in the benchmark exceeds a 5% rank-one rate. Teladoc's rank-one dominance is its strongest competitive asset, even if its overall recommendation coverage trails Sesame.

Recommendation-weighted winners. Sesame is the strongest recommendation-weighted winner, combining the highest valid recommendation coverage with the strongest net sentiment score of 0.82. Doctor on Demand is a close second, with strong recommendation coverage and a net sentiment score of 0.71. The combination of recommendation volume and positive framing gives these two brands the most commercially durable position in the benchmark.

Visible but under-recommended. Amwell is the category's clearest example of a visible-but-under-recommended brand. With 35.4% presence but only 20.2% recommendation coverage, the brand is being referenced in factual contexts and market overviews but is not being advanced in recommendation contexts. Its 8.2% top-three rate and 0.5% rank-one rate confirm that AI systems recognize Amwell without shortlisting it.

Strong recommendation quality despite lower raw visibility. MDLive achieves 26.3% recommendation coverage with a 40.4% presence rate and a net sentiment score of 0.72. While not a top-three contender, MDLive has established a credible second-tier position with positive framing. Its recommendation-to-presence conversion ratio is stronger than Amwell's, suggesting more effective source coverage relative to its footprint.

Present but commercially weak. LiveHealth Online, K Health, Lemonaid Health, and HealthTap all show presence rates below 10% and recommendation coverage below 7%. These brands are marginal players in AI-driven discovery, with occasional mentions but no meaningful shortlist presence in the consideration-stage cluster analyzed.

Cautionary visibility risk. Teladoc carries the highest negative visibility rate in the category at 0.54%, with three negative mentions across the 560 observations. While small in absolute terms, this pattern may indicate that some source material in the public evidence layer includes cautionary language or critical framing. In a trust-dependent category, even low-volume negative framing warrants monitoring.

Platform-specific patterns. Doctor on Demand performs strongest in Google AI Mode at 55.1% recommendation coverage. Sesame leads in Google AI Overviews at 57.6% coverage. Teladoc's highest platform performance is in Google AI Mode at 45% coverage. The variation across platforms suggests that source material is weighted differently by each AI system and that no single brand has achieved uniform dominance across all six platforms tested.

Why Visibility Is Not Enough

A brand can appear in AI answers and still fail to win the buyer shortlist. The benchmark demonstrates this distinction across multiple brands in the category, and the gap between presence and recommendation credit is the most important commercial insight in the data.

Amwell illustrates the problem most clearly. The brand appears in 35.4% of AI responses, meaning AI systems recognize it as a relevant market participant. But its valid recommendation coverage of 20.2% means that in the majority of those responses, Amwell is being mentioned without being advanced. The brand is cited, not recommended. It is present, not shortlisted. For a patient receiving an AI-generated answer, Amwell may appear as background context while Doctor on Demand or Sesame receives the actual recommendation.

The same principle applies at different scales across the full competitive set. Teladoc appears in 67.3% of responses but converts that presence into 41.8% recommendation coverage. MDLive appears in 40.4% of responses but converts to 26.3% coverage. Even the leaders do not convert every mention into a recommendation. The difference is that the leaders convert at rates high enough to maintain meaningful shortlist presence.

The commercial consequence is direct. Top-three placement corresponds to active patient consideration. Rank-one placement corresponds to the default choice for patients who follow the AI recommendation without comparison shopping further. Neutral mentions, comparison anchors, and factual references do not create that consideration. Brands that are merely mentioned risk being filtered out before they ever reach the patient's decision process, and the filtering happens at the AI response layer, not at the brand website.

The distinction also matters when evaluating net sentiment scores. A brand with a high presence rate and a low net sentiment score may be appearing in contexts where it is discussed critically, used as a cautionary example, or referenced in contexts that do not translate to recommendation credit. Sentiment and framing quality are separate from recommendation frequency, and both need to be tracked independently to understand competitive position.

The Citation Layer

AI systems build their responses from public source material, and the quality, consistency, and credibility of that material shapes which brands receive recommendation credit. The benchmark evidence suggests that several source types are influencing AI answers in the online doctors category.

Official brand sites provide the foundational entity information that AI systems use to understand what a provider offers, including service descriptions, provider credentials, accepted insurance, and pricing transparency. Brands with comprehensive, accurate, and consistently structured official content are easier for AI systems to describe with confidence and recommend with specificity. Thin or inconsistent official content may contribute to neutral framing or factual mention without recommendation credit.

Editorial reviews and comparison pages position brands against alternatives and help AI systems construct ranked shortlists. The platform-level differences observed in the benchmark, particularly Doctor on Demand's strength in Google AI Mode and Sesame's strength in Google AI Overviews, may reflect differences in how each platform retrieves and weights comparison-oriented editorial content.

Review platforms and community signals provide social proof that AI systems appear to weigh when assessing recommendation eligibility. In healthcare, this includes patient reviews, provider ratings, and community forum discussions where real-world experiences are documented. Positive, consistent review signals appear to support recommendation framing, while inconsistent or mixed signals may contribute to the neutral framing patterns observed in several mid-tier brands.

Regulatory and trust signals are particularly relevant in a healthcare context. AI systems appear to favor providers with clear licensing information, provider verification, credentialing documentation, and compliance transparency when deciding which services to recommend. Brands that surface these signals in searchable, public-facing content create a more credible evidence layer for AI systems to draw on.

Traditional search visibility remains a supporting factor. Brands with strong organic search footprints have more retrievable material available for AI systems to synthesize, which may help explain why certain narratives and framings appear consistently across multiple platforms. However, search visibility alone does not determine AI recommendation outcomes. The relationship between search visibility and AI recommendation credit should be treated as correlational, not causal, based on the evidence available in this benchmark.

What Brands Need to Fix

Weak valid recommendation coverage. Brands appearing frequently in AI responses but not being advanced as recommendations need to understand the gap between factual recognition and recommendation-level trust. The evidence suggests that factual recognition is necessary but not sufficient. Source material needs to convey recommendation-worthy attributes, not just categorical relevance.

Low top-three or rank-one presence. Brands that appear in AI responses but rarely in the first three positions are missing the most commercially valuable placement. Improving top-three and rank-one rates requires building source material that supports shortlist-quality positioning, including comparative advantages, trust signals, and third-party endorsements that AI systems can retrieve and synthesize.

Poor prompt-cluster coverage. The public benchmark covers the consideration-stage cluster, but the full LLM Authority Index report includes evaluation-stage comparison prompts and decision-stage pricing prompts. Brands absent from these higher-intent clusters are missing the buying moments where patient decisions are most actively forming. Prompt-cluster-specific gaps require targeted content and source coverage strategies.

Neutral or cautionary framing. Brands with high neutral visibility rates are being mentioned in contexts where AI systems are describing the market rather than recommending providers. Shifting from neutral to positive framing requires stronger endorsement signals in the public evidence layer, including editorial coverage that positions the brand as a recommended option rather than a market reference.

Thin source footprint. Brands with low presence and low recommendation coverage, particularly K Health, Lemonaid Health, and HealthTap, need to build a more substantial public evidence layer. AI systems have limited retrievable material to synthesize for these brands, which restricts both mention frequency and recommendation eligibility.

Inconsistent entity information. AI systems need consistent, accurate information about a brand's services, pricing, provider credentials, and service area. Inconsistent entity information across the public evidence layer can lead to neutral framing, incomplete descriptions, or exclusion from recommendation contexts where accuracy is required.

Weak third-party validation. The benchmark pattern suggests that AI systems weigh third-party signals, including editorial reviews, comparison rankings, and review platform signals, when deciding which brands to recommend. Brands without strong third-party coverage are at a structural disadvantage relative to brands like Sesame and Doctor on Demand, which appear to have deeper editorial and review support.

Limited citation architecture. Brands that are visible but not recommended may have a public evidence layer that supports factual reference but not recommendation-level trust. The priority is building citation architecture that supports positive, shortlist-quality framing across source types, including editorial, review, regulatory, comparison, and official content layers.

How CiteWorks Studio Helps

1. Map AI recommendation visibility. Track prompts, platforms, company presence, valid recommendations, top-three and rank-one performance, framing quality, and citation sources to establish a precise recommendation-stage baseline.
2. Identify the sources shaping AI answers. Find the editorial, review, forum, government, directory, owned, search-visible, and backlink-supported sources that are influencing brand framing and recommendation credit across AI platforms.
3. Build the citation architecture plan. Strengthen the public evidence layer so AI systems have more accurate, consistent, and persuasive source material to synthesize when constructing recommendations in high-intent prompt clusters.

Commercial Takeaway

AI-led discovery is changing where patient shortlists are formed. When a patient asks an AI assistant which online doctor service to use, the response effectively creates the consideration set. Brands that appear in that shortlist, particularly in top-three and rank-one positions, gain disproportionate access to patient demand at the highest-intent moment in the discovery journey. Brands that are merely mentioned risk being filtered out before they reach the patient's consideration.

The benchmark shows that brands can lose recommendation-stage visibility even when they are visible in AI answers. Amwell's 35.4% presence rate means the brand is being seen. Its 20.2% recommendation coverage means it is not being chosen. Competitors like Doctor on Demand and Sesame are intercepting demand in the highest-intent prompt clusters, and the concentration effect is self-reinforcing. Recommended brands generate more review content, more comparison coverage, more editorial authority, and more official brand signals, all of which feed back into stronger AI recommendation performance over time.

Traditional search and source visibility still matter because they contribute to the public evidence layer that AI systems draw from. But the strategic opportunity is improving recommendation-stage visibility, not merely chasing presence. Brands that build comprehensive citation architectures across official content, comparison sources, review platforms, and trust signals will capture disproportionate share of patient demand as AI platforms become a primary discovery mechanism for healthcare services.

See How AI Is Recommending Your Brand

The benchmark shows where the market stands. Every brand has a different recommendation profile, and the gaps in your specific profile determine where competitors are intercepting your demand.

CiteWorks Studio can show where your brand appears in AI-generated responses, where competitors are recommended instead, which prompts carry the most commercial risk, which sources are shaping AI answers about your category, and what needs to change to improve recommendation-stage visibility.

Request an AI Visibility Audit, an AI Market Discovery Profile, or a Citation Architecture Review to map your brand's AI recommendation footprint and identify the highest-priority remediation opportunities.

Benchmark Source

This analysis is based on the 2026 AI Discovery Index for Online Doctors, published by LLM Authority Index. Read the full benchmark report at the LLM Authority Index website.