Supplements AI Visibility Index 2026

Thorne leads everyday wellness and dietary supplement AI visibility in this measurement, appearing in 35.6% of 1,709 returned answers. Garden of Life follows at 31.4%, and NOW ranks third at 26.3%. The measured results do not support a simple claim that certification references alone dictate brand recommendation share: USP and NSF are frequently named as standards, while the appearance leaders are brands.

This flagship roll-up covers multivitamins, protein and protein powder, probiotics and gut health, sleep products involving melatonin or magnesium, immune support, and nootropics or cognitive products. The results describe these 144 measured prompts on September 30, 2026. They do not measure all supplement searches, consumer demand, product quality, or safety.

The cards summarize the study's appearance and citation measurements. Appearance rates use returned answers, while citation rates and source penetration use answers with retrieval evidence.

Which brands lead the overall leaderboard?

Thorne ranks first with a 35.6% appearance rate and a 95% interval of 33.3% to 37.9%. Garden of Life ranks second at 31.4%, followed by NOW at 26.3%, Nature Made at 23.1%, and Pure Encapsulations at 19.8%.

These are appearance results, using 1,709 returned answers. A brand can appear in more than one answer, so appearance rates do not add up to a fixed total.

Appearance rate is the share of the 1709 returned AI answers in which each brand was named.

The rank 1 interval overlaps the intervals for Garden of Life, NOW, Nature Made, and Pure Encapsulations. Within the margin of this sample, the study cannot claim Thorne is more visible than those brands. The ranking records the measured point estimates, while the intervals show the uncertainty around them.

How do the leading brands appear across engines?

Thorne

35.6% appearance rate, rank 1, with a 33.3% to 37.9% interval. Thorne appeared in Perplexity at 38.9%, ChatGPT at 53.5%, Gemini at 28.9%, and Claude at 21.8%, so every measured engine named it. Its average first mention position was 3.4. Its owned domain was cited in 6.3% of answers with retrieval evidence.

Garden of Life

31.4% appearance rate, rank 2, with a 29.2% to 33.6% interval. Garden of Life appeared in Perplexity at 42.1%, ChatGPT at 57.4%, Gemini at 17.8%, and Claude at 9.3%. Its average first mention position was 6.1. Its owned domain was cited in 3.3% of answers with retrieval evidence.

NOW

26.3% appearance rate, rank 3, with a 24.2% to 28.4% interval. NOW appeared in Perplexity at 26.6%, ChatGPT at 53%, Gemini at 21.1%, and Claude at 5.6%. Its average first mention position was 6.7. Its owned domain was cited in 0.1% of answers with retrieval evidence.

Nature Made

23.1% appearance rate, rank 4, with a 21.2% to 25.2% interval. Nature Made appeared in Perplexity at 35%, ChatGPT at 34.9%, Gemini at 15.7%, and Claude at 7.4%. Its average first mention position was 5. Its owned domain was cited in 4.8% of answers with retrieval evidence.

Pure Encapsulations

19.8% appearance rate, rank 5, with an 18% to 21.7% interval. Pure Encapsulations appeared in Perplexity at 16%, ChatGPT at 33.4%, Gemini at 18.8%, and Claude at 11.6%. Its average first mention position was 4.3. Its owned domain was cited in 0.2% of answers with retrieval evidence.

Life Extension

17.8% appearance rate, rank 6, with a 16.1% to 19.7% interval. Life Extension appeared in Perplexity at 21.3%, ChatGPT at 33.4%, Gemini at 13.4%, and Claude at 3.9%. Its average first mention position was 5.2. Its owned domain was cited in 9.2% of answers with retrieval evidence.

Ritual

15.3% appearance rate, rank 7, with a 13.6% to 17.1% interval. Ritual appeared in Perplexity at 21.5%, ChatGPT at 14.3%, Gemini at 13.4%, and Claude at 11.8%. Its average first mention position was 4.5. Its owned domain was cited in 1.1% of answers with retrieval evidence.

Transparent Labs

13% appearance rate, rank 8, with an 11.5% to 14.7% interval. Transparent Labs appeared in Perplexity at 21.5%, ChatGPT at 9.2%, Gemini at 10%, and Claude at 11.3%. Its average first mention position was 3.3. Its owned domain was cited in 1.8% of answers with retrieval evidence.

Optimum Nutrition

12.1% appearance rate, rank 9, with a 10.6% to 13.7% interval. Optimum Nutrition appeared in Perplexity at 14.8%, ChatGPT at 16.5%, Gemini at 8.1%, and Claude at 9%. Its average first mention position was 3.7. Its owned domain was cited in 1.6% of answers with retrieval evidence.

Culturelle

9.7% appearance rate, rank 10, with an 8.4% to 11.2% interval. Culturelle appeared in Perplexity at 13.9%, ChatGPT at 13.1%, Gemini at 4.6%, and Claude at 7.4%. Its average first mention position was 3.8. Its owned domain was cited in 1.5% of answers with retrieval evidence.

Where do the engines disagree?

The engines disagree most sharply on several brands, and these differences may reflect model and retrieval-tool design rather than a brand's overall visibility. Perplexity returned 432 answers, ChatGPT returned 413, Gemini returned 432, and Claude returned 432. Each engine rate uses that engine's own returned-answer total.

Each cell uses that engine's own answer count as its denominator, so engines can be compared row by row.

ChatGPT named Garden of Life in 57.4% of its 413 answers and NOW in 53%, compared with 9.3% and 5.6% in Claude's 432 answers. Thorne also had a wide spread, from 53.5% in ChatGPT to 21.8% in Claude. Nature Made was closer across Perplexity and ChatGPT at 35% and 34.9%, then fell to 15.7% in Gemini and 7.4% in Claude.

Gemini repeatedly named standards and sources that other engines rarely or never named. USP appeared in 28% of Gemini answers, 7.2% of Perplexity answers, 0.5% of ChatGPT answers, and 0% of Claude answers. NSF appeared in 14.6% of Gemini answers, 6.7% of Perplexity answers, 0.5% of ChatGPT answers, and 0% of Claude answers. National Institutes of Health appeared in 16% of Gemini answers but 0% in Perplexity and Claude.

Some brands were concentrated in a single system. Nature's Bounty appeared in 18.3% of Perplexity answers and 0% of Claude answers. Jarrow Formulas appeared in 19.4% of ChatGPT answers but 0.9% of Claude answers. Myprotein appeared in 8.8% of Perplexity answers and 8.5% of ChatGPT answers, then 0% of Claude answers. These gaps mean a single-engine check can produce a misleading picture of category visibility.

Which brands lead each demand cluster?

The demand-cluster results show where brands are most often named within the measured supplement questions. General selection questions and more specific need-based questions measure different forms of demand, so the cluster leaders should be read separately rather than as one universal ranking.

The chart separates broad buyer-intent questions from sub-category and specific-need questions. A brand that leads a narrow cluster may have a focused strength without leading the overall roll-up.

Which domains receive citations and where do answers come from?

Owned-domain citation rates use 1,707 answers with retrieval evidence. National Institutes of Health had the highest listed owned-domain citation rate at 22%, followed by Life Extension at 9.2%, Mind Lab Pro at 6.8%, Thorne at 6.3%, Seed at 4.9%, and Nature Made at 4.8%.

Share of the 1707 answers with retrieval evidence that cited a URL on the brand's own website.

Appearance and owned-domain citation are different measurements. NOW ranked third in appearance at 26.3% but had a 0.1% owned-domain citation rate. Life Extension ranked sixth in appearance at 17.8% and had a 9.2% owned-domain citation rate. The data therefore does not support treating appearance as a proxy for a brand site's citation frequency.

Third-party publishers and retailers appear far more often in the source tally than individual supplement brand sites. Healthline was cited in 56.1% of answers with retrieval evidence, Forbes in 35.1%, Fortune in 32.6%, ConsumerLab in 31%, and Garage Gym Reviews in 20.5%.

Publishers cited as evidence, counted once per answer rather than once per link.

The cited-source pattern shows that answer generation drew heavily on health publishers, testing resources, review sites, and retailers. ConsumerLab appeared in 31% of retrieval-backed answers, while Amazon appeared in 18%, Walmart in 11.7%, GNC in 11.3%, and Target in 11.2%.

Who is winning and who is at risk?

Brands with the strongest measured positions

Thorne leads the appearance tally at 35.6%, appears across all four engines, and averages a 3.4 first mention position. Garden of Life has a 31.4% appearance rate and leads the measured ChatGPT rate at 57.4% among the top brands. Nature Made combines a 23.1% appearance rate with a 4.8% owned-domain citation rate. Life Extension combines a 17.8% appearance rate with a 9.2% owned-domain citation rate. Transparent Labs has a 13% appearance rate and a 3.3 average first mention position.

Brands with measured exposure risks

NOW has a 26.3% appearance rate but a 0.1% owned-domain citation rate. Garden of Life has a 31.4% appearance rate but a 9.3% Claude rate, compared with 57.4% in ChatGPT. Nature's Bounty appears in 18.3% of Perplexity answers but 0% of Claude answers. Myprotein appears in 8.8% of Perplexity answers and 0% of Claude answers. Airborne has a 2.9% appearance rate and a 0% owned-domain citation rate.

What structural patterns does the tally show?

The top of the appearance ranking is close enough for caution. Thorne's 33.3% to 37.9% interval overlaps Garden of Life, NOW, Nature Made, and Pure Encapsulations, so the sample does not establish a clear separation among those positions. The top brands are visible across four engines, but their engine-level rates diverge sharply.

Third-party sources carry a large share of retrieval-backed citations. Healthline, Forbes, Fortune, ConsumerLab, Garage Gym Reviews, Innerbody, Amazon, Medical News Today, Everyday Health, and U.S. News Health all appear in the cited-source tally. The highest listed brand-owned domain rate is Life Extension at 9.2%, while Healthline's answer penetration is 56.1%.

The study also recorded near-zero appearance for several measured brands. Calm appeared in 0.1% of returned answers, Moon Juice in 0.5%, and Physician's Choice in 0.6%. Those results are limited to these prompts and do not measure product quality or consumer demand.

What changed during the measurement window?

The tally does not measure change over time. It measures one cross-segment roll-up, so it cannot identify risers, fallers, or category shifts during the measurement window.

What should brands do with these measured gaps?

• Audit engine-specific coverage. Garden of Life appeared in 57.4% of ChatGPT answers and 9.3% of Claude answers. Re-measure the same engine split after publishing materials that answer the product-selection questions where the gap appears.
• Build citation-ready product pages. NOW appeared in 26.3% of returned answers but its owned domain was cited in 0.1% of retrieval-backed answers. Re-measure owned-domain citation rate against the same 1,707-answer retrieval-evidence denominator.
• Publish pages that state product purpose, ingredients, and selection constraints in plain language. The source tally includes Healthline at 56.1% and ConsumerLab at 31%, showing that answers often cite explanatory third-party pages. Re-measure whether the brand domain appears alongside those sources.
• Test content across every engine. Nature's Bounty appeared in 18.3% of Perplexity answers and 0% of Claude answers. Re-measure engine-level appearance after changes rather than using one engine as the scorecard.
• Separate broad-category pages from focused need pages. The cluster chart distinguishes general selection questions from specific-need questions. Re-measure appearance by cluster to identify whether a page expands broad discovery or only a narrow product use case.
• Make verification information easy to locate and describe precisely. USP appeared in 28% of Gemini answers and NSF in 14.6%, while the leading brands varied by engine. Re-measure whether a brand's owned domain is cited more often when it explains its testing or verification information without unsupported claims.

What does this mean for supplement brands?

Interpretation: the measurement suggests that broad brand visibility and source citation are separate tasks. Brands can be named often without their own site being cited often, and a strong result in one engine may not travel to another. The practical implication is to track appearance, first-mention position, engine spread, and owned-domain citations as separate measures.

Interpretation: third-party health publishers, testing resources, review sites, and retailers are a major part of the source mix in these answers. A brand's own pages need to answer the same concrete selection questions if the brand wants to test whether its site can earn more retrieval-backed citations.

Where can I read each category report?

Each of the six categories in this index has its own full report, built only from that category's measured answers, with its own brand leaderboard, engine comparison and findings.

• Multivitamins AI Visibility Index 2026: Thorne was the most-named brand, in 56.9% of 288 measured answers.
• Protein Powder AI Visibility Index 2026: Optimum Nutrition was the most-named brand, in 64.2% of 288 measured answers.
• Probiotics & Gut Health AI Visibility Index 2026: Culturelle was the most-named brand, in 56.4% of 273 measured answers.
• Sleep Supplements AI Visibility Index 2026: Thorne was the most-named brand, in 52.1% of 288 measured answers.
• Immune Support AI Visibility Index 2026: Thorne was the most-named brand, in 47.6% of 286 measured answers.
• Nootropics & Cognitive AI Visibility Index 2026: Mind Lab Pro was the most-named brand, in 37.8% of 286 measured answers.

Charts and data tables

These are not brands in the category, so they are measured the same way but kept out of the brand leaderboard.

Methodology

This study measured 144 prompts across Perplexity, ChatGPT, Gemini, and Claude, with 3 trials per prompt per engine. Appearance rates use 1,709 returned answers. Citation and cited-source rates use 1,707 answers with retrieval evidence. The study recorded 1,717 attempted answers, 10 answers without retrieval evidence, and no failed answers.

How this study was measured

• Sampling frame fixed first: 144 prompts allocated across the demand clusters before anything was asked.
• Same questions put to every engine: Perplexity (sonar-pro), ChatGPT (OpenAI), Gemini (Google), Claude (Anthropic), 3 trial(s) each.
• 1709 answers returned of 1717 attempted, 1707 with retrieval evidence.
• Product lines rolled up to parent brands using a registry fixed before scoring.
• Appearance rate reported with a 95% Wilson interval; citation metrics counted only on answers with evidence.

The findings are a cross-segment roll-up from 1,717 measured answers across six segment studies. The measurement identifies names and cited domains in these prompts on this date. It does not establish product quality, medical efficacy, safety, consumer demand, or visibility beyond the measured prompt set.

Methodology appendix

This appendix is generated directly from the measurement record for this study. It is included so any brand named here can reproduce or challenge the result.

What was measured

• Category: Everyday Wellness & Dietary Supplements
• Measurement date: 2026-09-30
• Depth: flagship — 144 prompts, 3 trial(s) per prompt per engine, 4 engines
• Answers attempted: 1717
• Answers returned (appearance denominator): 1709
• Answers with retrieval evidence (citation denominator): 1707 (99.4% of attempted)
• Answers that failed outright: 0
• Analysis version: study-methodology-v2 · entity registry: entity-registry-2026-08-27

Claim limit: Cross-segment roll-up: every figure in this hub is tallied from the 1717 measured answers of its 6 segment studies. No figure here is estimated, and no category is ranked that was not measured in one of those segments.

How the prompts were sampled

Results are not strictly comparable across engines: each engine uses a different model, a different retrieval tool and a different answer format, so a gap between two engines can reflect system design as much as brand visibility.

How each metric is defined

• Appearance rate — the share of returned answers that named the brand. Measured on every answer the engine returned, whether or not it searched the web, so an engine's retrieval behaviour does not silently change the denominator.
• Macro appearance rate — the mean of the per-prompt appearance rates, so each prompt counts once regardless of how many engines answered it.
• 95% interval — Wilson score interval on the appearance rate. Where two brands' intervals overlap, this study does not support a claim that one is more visible than the other.
• Owned-domain citation rate — the share of answers with retrieval evidence that cited a URL on the brand's own domain. This is a narrow measure of the brand's own site being used as a source; it is not a measure of total mentions.
• Share of all citations — the brand's own domain as a share of all attributed publisher citations in the study.
• Not measured — the brand's official domain could not be established, so no citation result exists. This is never reported as 0%.
• 0% — the metric could be evaluated and the event did not occur.
• Average position — mean position of the brand's first mention in the answer's ordered list of named brands.
• Cross-engine consistency — the share of engines that named the brand at least once.
• Repeat consistency — the share of repeated prompt/engine cells where the brand appeared in every trial. Lower values mean the engine's answer is unstable rather than the brand being absent.

Evidence rules

• An answer counts as having retrieval evidence only when the engine returned a retrieval receipt (search annotations or grounding records). A URL the model merely typed into its prose is not treated as evidence that it searched.
• Sources are attributed to the publisher behind each retrieval receipt. Intermediary and redirect hosts are never reported as publishers.
• Publisher identity could not be established for 0 retrieval receipt(s); those are disclosed here rather than attributed or dropped.
• Source counts are answer-level: a publisher counts once per answer in which it was cited, not once per URL.

Entity handling

Product lines and model names roll up to their parent brand for brand-level metrics, using a registry fixed before scoring. Nothing is merged on the basis of this study's own counts alone.

• Ghost counted under Brooks (product_line)

40 further name(s) picked up in the answers were not confidently resolvable to a brand in this category and are held for human classification before the next study; they carry no metrics here.

Limits of these findings

• This is a point-in-time measurement of what the named systems returned for these exact prompts on 2026-09-30. Engine behaviour changes without notice.
• The prompt set is a stratified sample of category demand, not a census. Results describe these prompts and should not be read as representative of all searches in the category.
• Where intervals overlap, differences between brands are not statistically distinguishable at this sample size.
• A brand not listed here was not measured as absent — it simply was not named in these answers.
• Appearance is a visibility measure, not an endorsement, a quality judgement or a market-share estimate.
• 1 of 52 ranked brand(s) had no official domain established before scoring, so they carry appearance results only and are excluded from citation metrics and source tables rather than being reported as zero.