Immune Support AI Visibility Index 2026

Immune support supplement visibility varied by engine in this measurement. Thorne ranked first with a 47.6% appearance rate among answers returned, while NOW Foods ranked second at 44.8%. The difference is within the margin of this sample because their 95% intervals overlap. The measurement covered Vitamin C, Zinc, Elderberry, Vitamin D, and mushroom formulas.

Immune Support is the sole focus of this standalone report in the Supplements AI Visibility Index 2026. Multivitamins, Protein / Protein Powder, Probiotics / Gut Health, Sleep (melatonin / magnesium), and Nootropics / Cognitive are outside its scope. When an answer includes supporting links, a brand mentioned frequently but seldom cited can face a practical visibility gap.

The results describe the measured prompts on 2026-09-30, not all immune support searches or consumer demand.

Which immune support brands lead the leaderboard?

Thorne led the appearance leaderboard with a 47.6% rate among 286 answers returned. NOW Foods followed at 44.8%. Nature Made and Garden of Life were tied at 39.2%, followed by Pure Encapsulations at 28.7% and Life Extension at 28%.

Among the brands named in the client brief, Emergen-C ranked eighth at 19.9%, Airborne ranked eleventh at 17.5%, Sambucol ranked twelfth at 15.7%, and Host Defense ranked twenty-sixth at 3.8%.

Immune support appearance leaderboard

Appearance rate shows the proportion of the 286 returned AI answers that named each brand.

Thorne's 95% interval of 41.8% to 53.3% overlaps NOW Foods at 39.1% to 50.5%, Nature Made at 33.7% to 44.9%, Garden of Life at 33.7% to 44.9%, and Pure Encapsulations at 23.7% to 34.2%. The observed differences among these brands are within the margin of this sample. Nature Made and Garden of Life have identical appearance rates and intervals.

While the top brands appear across a wide range of engines, their engine-by-engine patterns vary considerably. The profiles below calculate appearance rates from answers returned. Owned-domain citation rates are based on answers with retrieval evidence.

Where do the engines disagree?

The engines disagree most on which brands deserve repeated naming. Perplexity (sonar-pro), ChatGPT (OpenAI), Gemini (Google), and Claude (Anthropic) use different models and retrieval tools, so an engine gap can reflect system design rather than a brand's underlying visibility.

Engine-by-brand appearance rates

Every cell is calculated against that engine's own answer count, allowing row-by-row engine comparisons.

ChatGPT (OpenAI) concentrated heavily on several leading brands across 70 returned answers: NOW Foods at 84.3%, Garden of Life at 80%, Thorne at 74.3%, Life Extension at 54.3%, and both Pure Encapsulations and Nordic Naturals at 50%. Claude (Anthropic) named the same brands far less often, including NOW Foods at 19.4%, Garden of Life at 4.2%, Life Extension at 5.6%, and Nordic Naturals at 5.6% across 72 returned answers.

Perplexity (sonar-pro) favored Nature Made at 66.7% of 72 returned answers and Nature's Bounty at 44.4%. Claude (Anthropic) named Nature Made in 9.7% and did not name Nature's Bounty in its 72 returned answers. GNC appeared in 43.1% of Perplexity's answers but 4.3% of ChatGPT's and 0% of Gemini's and Claude's answers. New Chapter had a similar Perplexity pattern: 26.4%, versus 2.9%, 6.9%, and 5.6% in the other engines.

Gemini (Google) produced a distinct source and certification pattern. It named ConsumerLab in 54.2% of 72 returned answers, USP in 55.6%, United States Pharmacopeia in 25%, NSF in 25%, and National Institutes of Health in 31.9%. Claude (Anthropic) named none of ConsumerLab, USP, NSF, or National Institutes of Health in its 72 returned answers. These entities are sources or certification bodies, not supplement brands.

Nature's Way appears twice in the tally under typographic variants. Nature's Way appeared in 15.7% of returned answers, while Nature's Way with a straight apostrophe appeared in 8.4%. The variants have the same domain, naturesway.com, and their engine patterns differ. This naming split is part of the observed tally and limits any combined interpretation.

Which demand clusters favor which brands?

The demand-cluster results show which brands lead particular types of immune support questions within the measured prompt set. The chart separates the cluster groups, so general buyer-intent patterns should not be read as the same kind of evidence as sub-category and specific-need patterns.

Brand leaders by demand cluster

Here, the answers are regrouped according to the demand cluster from which each prompt was sampled; each cluster rate uses its own answer count.

For general buyer-intent questions, the chart identifies the brands that recur when an answer names products or makers. For sub-category and specific-need questions, it identifies where a brand's measured visibility is tied to Vitamin C, Zinc, Elderberry, Vitamin D, or mushroom formulas. A brand can lead one cluster without leading the full leaderboard because the prompts test different question types.

Which sites receive citations and which sources supply answers?

Third-party sources received more prominent citation support than most supplement makers' own sites. Citation rates below use the 286 answers with retrieval evidence, not the 286 returned answers used for appearance.

Owned-domain citation rates

This is the share of the 286 answers with retrieval evidence that cited a URL from the brand's own website.

Among citation-eligible supplement brands, GNC had a 19.2% owned-domain citation rate, Life Extension had 12.2%, Thorne had 9.4%, New Chapter had 8%, and Nature Made had 7.7%. Airborne had a 0% owned-domain citation rate despite a 17.5% appearance rate among answers returned. These are different measures: appearance records a brand name in an answer, while citation measures whether retrieval evidence cited the brand's own domain.

Most-cited source domains

Evidence publishers are counted once in each answer, rather than once for every link.

ConsumerLab.com had 60.5% answer penetration among answers with retrieval evidence. Healthline.com had 36.4%, health.usnews.com had 25.2%, amazon.com had 19.6%, and gnc.com and walmart.com each had 19.2%. The top source pattern includes health publishers, retailers, and testing or comparison material. It does not support a claim that any one source type controls all immune support answers.

Who is winning and who is at risk?

Who is winning?

Thorne is winning on broad measured appearance: it ranked first at 47.6%, appeared in all four engines, and had an average first-mention position of 2.7. NOW Foods ranked second at 44.8% and reached 84.3% in ChatGPT's 70 returned answers. Nature Made and Garden of Life shared a 39.2% appearance rate, while their engine profiles differed. Life Extension paired a 28% appearance rate with a 12.2% owned-domain citation rate among answers with retrieval evidence.

Who is at risk?

Airborne is at risk of a citation gap: it appeared in 17.5% of returned answers but had a 0% owned-domain citation rate among answers with retrieval evidence. Host Defense ranked twenty-sixth at 3.8%, with a 95% interval of 2.2% to 6.8%, and its rates stayed between 2.8% and 4.3% across engines. Emergen-C ranked eighth at 19.9%, but its average first-mention position was 7.1 and its owned-domain citation rate was 1%. Sambucol ranked twelfth at 15.7% with a 1.4% owned-domain citation rate.

What structural pattern does the tally show?

The top of the appearance leaderboard is close rather than decisive. Thorne's 41.8% to 53.3% interval overlaps the intervals for NOW Foods, Nature Made, Garden of Life, and Pure Encapsulations, so the sample does not establish that Thorne is more visible than those brands. The engines also do not converge on one stable brand set. ChatGPT repeatedly named NOW Foods and Garden of Life, Perplexity repeatedly named Nature Made, and Gemini repeatedly named certification and testing entities.

Owned-domain citation patterns are separate from appearance patterns. ConsumerLab had a 60.8% owned-domain citation rate, National Institutes of Health had 30.8%, and GNC had 19.2%, while several frequently named supplement brands had much lower owned-domain citation rates. The source list also places ConsumerLab.com, Healthline.com, health.usnews.com, amazon.com, gnc.com, and walmart.com among the cited domains. The measured answer set therefore draws support from publishers, retailers, testing sources, government material, and brand sites.

Among the brief's named brands, Host Defense was near the bottom of the measured leaderboard at 3.8%. Emergen-C, Airborne, and Sambucol appeared more often, but each had an owned-domain citation rate at or below 1.4% among answers with retrieval evidence.

What changed during the measurement window?

The data contains no time-series comparison, so it does not show a change during the measurement window. It records repeat generations across the measured prompt set and shows engine-level divergence, including Gemini's higher appearance rates for ConsumerLab, USP, NSF, and National Institutes of Health, and Perplexity's higher rates for GNC and New Chapter.

What should immune support brands fix first?

• Airborne should inspect the pages used for its immune support products because it had a 17.5% appearance rate but a 0% owned-domain citation rate. Re-measure its owned-domain citation rate across the same 24 prompts and four engines.
• Emergen-C should strengthen material that can support earlier answer placement because its average first-mention position was 7.1. Re-measure average first-mention position alongside its 19.9% appearance rate.
• Host Defense should build coverage around the specific mushroom-formula questions included in this segment because it ranked twenty-sixth at 3.8%. Re-measure its cluster results and overall appearance rate.
• Sambucol should test whether its elderberry pages can earn more retrieved support because its owned-domain citation rate was 1.4%. Re-measure citations among answers with retrieval evidence, not among all returned answers.
• NOW Foods should reduce its engine dependence because its appearance rate was 84.3% in ChatGPT but 19.4% in Claude. Re-measure each engine separately rather than relying on one combined rate.
• Nature Made should investigate its Perplexity and Claude gap because it appeared in 66.7% of Perplexity's answers and 9.7% of Claude's. Re-measure the same brand across each engine and preserve the engine-specific denominator.
• Nature's Way should standardize how the brand is named in pages and supporting material because the tally contains both Nature's Way and Nature's Way variants under the same domain. Re-measure whether those variants remain split in future answer tallies.

What does this mean for immune support brands?

Interpretation: Immune support brand visibility is not one contest with one answer. The measurement shows different answer sets by engine, while cited support often comes from testing sources, health publishers, retailers, government material, and selected brand sites. Brand teams can separate two jobs: earn brand mentions in the answers returned, and earn citations to the pages that support those answers.

For seasonal prevention, elderberry, zinc, Vitamin D, Vitamin C, and mushroom formula questions, the next measurement should preserve both distinctions. A rise in appearance without a rise in owned-domain citations would mean a different outcome than a rise in citations without broad engine coverage.

How was this immune support study measured?

This flagship study tested 24 stratified prompts three times in Perplexity (sonar-pro), ChatGPT (OpenAI), Gemini (Google), and Claude (Anthropic). It attempted 288 answers and returned 286. Appearance rates use the 286 returned answers. Citation and source rates use the 286 answers with retrieval evidence. Results are limited to the measured prompts and repeat generations on 2026-09-30.

Measurement summary

How this study was measured

• The sampling frame came first: 24 prompts were assigned across the demand clusters before any questions were asked.
• Every engine received the same questions: Perplexity (sonar-pro), ChatGPT (OpenAI), Gemini (Google), Claude (Anthropic), 3 trial(s) each.
• Of 288 attempted, 286 answers were returned, and 286 had retrieval evidence.
• Product lines were consolidated under parent brands through a registry set before scoring.
• Appearance rate uses a 95% Wilson interval; citation metrics include only answers with evidence.

The study does not measure sales, clinical effectiveness, product quality, consumer demand, or the full universe of immune support searches. It measures which named entities and cited domains appeared in this defined answer set.

Charts and data tables

Although measured in the same way, these are not brands in the category and are excluded from the brand leaderboard.

Methodology appendix

This appendix comes directly from the study's measurement record. It is provided so any brand named here can reproduce or contest the result.

What was measured

• Category: Immune Support
• Measurement date: 2026-09-30
• Depth: flagship — 24 prompts, 3 trial(s) per prompt per engine, 4 engines
• Answers attempted: 288
• Answers returned (appearance denominator): 286
• Answers with retrieval evidence (citation denominator): 286 (99.3% of attempted)
• Answers that failed outright: 0
• Analysis version: study-methodology-v2 · entity registry: entity-registry-2026-08-27

Claim limit: Flagship depth: 24 stratified prompts were measured three times per engine, making both between-prompt and repeat-generation variation observable.

How the prompts were sampled

The sampling frame is fixed: 14 buyer-intent need(s) from the client's brief are asked verbatim, one prompt each, grouped under: Immune Support; then , then a documented set of generic commercial demand clusters (general best, everyday mainstream choice, price/value, primary use case, beginner, retail channel, gifting/occasion, specific need, comparison, trust, ownership attribution, quality tier, durability, specialist, emerging) receives the remaining allocation round-robin. The prompt writer provides wording only within these cells and does not determine the category's demand structure.

Engine results are not strictly comparable: each uses a different model, retrieval tool and answer format, meaning a gap between two engines may reflect system design as much as brand visibility.

How each metric is defined

• Appearance rate — the proportion of returned answers that named the brand. It is measured across every answer an engine returned, whether or not the engine searched the web, so retrieval behaviour does not quietly alter the denominator.
• Macro appearance rate — the average of per-prompt appearance rates, giving every prompt one count regardless of the number of engines that answered it.
• 95% interval — the Wilson score interval for appearance rate. When two brands' intervals overlap, this study does not support saying that either is more visible.
• Owned-domain citation rate — the proportion of answers with retrieval evidence that cited a URL on the brand's own domain. It narrowly measures use of the brand's own site as a source, not total mentions.
• Share of all citations — the proportion of all attributed publisher citations in the study accounted for by the brand's own domain.
• Not measured — no citation result exists because the brand's official domain could not be established. It is never presented as 0%.
• 0% — the metric was assessable, but the event did not happen.
• Average position — the average position of the brand's first mention within the answer's ordered list of named brands.
• Cross-engine consistency — the proportion of engines that named the brand at least once.
• Repeat consistency — the proportion of repeated prompt/engine cells in which the brand appeared in every trial. Lower figures indicate an unstable engine answer, rather than a brand's absence.

Evidence rules

• Retrieval evidence is counted only when the engine returned a retrieval receipt (search annotations or grounding records). A URL merely typed into model prose does not establish that it searched.
• Sources are assigned to the publisher behind each retrieval receipt. Intermediary and redirect hosts are not reported as publishers.
• Publisher identity could not be established for 0 retrieval receipt(s); these are disclosed here instead of being attributed or omitted.
• Source counts are calculated at answer level: a publisher is counted once per answer that cited it, not once per URL.

Entity handling

For brand-level metrics, product lines and model names are assigned to their parent brand under a registry fixed before scoring. No grouping is made solely from this study's own counts.

No roll-ups were applied in this study.

The answers produced 40 further name(s) that could not be confidently resolved to a brand in this category. They are awaiting human classification before the next study and have no metrics here.

Limits of these findings

• This is a point-in-time record of what the named systems returned to these exact prompts on 2026-09-30. Engine behaviour can change without notice.
• The prompt set is a stratified sample of category demand rather than a census. Findings apply to these prompts and should not be treated as representative of all category searches.
• At this sample size, brands with overlapping intervals cannot be statistically distinguished.
• A brand absent from this list was not measured as absent — it was simply not named in these answers.
• Appearance measures visibility, not endorsement, quality judgement or market-share estimate.
• Before scoring, no official domain was established for 2 of 26 ranked brand(s). They therefore have appearance results only and are excluded from citation metrics and source tables, rather than reported as zero.