AI Beauty Radar #2 · June 2026 issue · 150 prompts

SPF50 sun care: what ChatGPT and Gemini recommend

Second issue of the AI Beauty Radar: 150 prompts on SPF50 sun care, tested on ChatGPT and Gemini across five shopping universes (86 ranked prompts, 57 advice prompts analysed separately, 7 brand-review prompts), with more recent models than issue #1.

Analysis by MYA, Global Search agency (SEO + GEO) · V.M.R. methodology (Visible, Mentioned, Recommended) · Google AI Overviews since March 2025

150

prompts tested per AI: 86 ranked, 57 advice, 7 reviews

5

shopping universes, on ChatGPT and Gemini

86%

of ChatGPT's advice answers name no brand (2% for Gemini)

73%

of Gemini's shopping answers name a retailer (2% for ChatGPT)

Tests run on 1 June 2026 on ChatGPT (GPT-5.5 Instant) and Gemini (Gemini 3.5 Flash), using French-language prompts, with a new conversation for each prompt and only the first response analysed. Presence rate measures the share of prompts in which a brand is mentioned or recommended. 150 prompts were tested per AI, including 86 used in the rankings. “First mention” indicates the share of responses in which the brand is named first. The results provide an indication of brand visibility at the time of testing.

Each Radar is a snapshot of brand visibility in AI at its test date. As models evolve regularly, results are not strictly comparable from one issue to the next.

Ranking: Face sun cream

Base: 17 product-choice or problem prompts ranked, 13 advice prompts analysed separately.

How users search. They describe a feeling, rarely a brand: “doesn't ball up under foundation”, “goes shiny after 2 hours”, “I hate the greasy feel”. Nearly one prompt in three is about make-up (pilling, reapplication, waiting time before foundation). The others are about sensitive skin, fragrance-free formulas and white marks on tan or dark skin.

Bar and first figure: presence rate (Mentioned). Light part of the bar and greyed figure: first mention (recommendation signal).

ChatGPT
  1. 1La Roche-Posay100%41.2%
  2. 2Eucerin82.4%52.9%
  3. 3Bioderma82.4%0%
  4. 4Avène64.7%5.9%
  5. 5SVR47.1%0%
  6. 6Isdin17.6%0%
  7. 7Vichy17.6%0%
Gemini
  1. 1La Roche-Posay100%64.7%
  2. 2SVR82.4%11.8%
  3. 3Isdin70.6%0%
  4. 4Bioderma52.9%0%
  5. 5Avène47.1%0%
  6. 6Eucerin41.2%17.6%
  7. 7Garnier23.5%0%
  8. 8A-Derma11.8%0%
  9. 9Make Up For Ever11.8%0%
  10. 10Nivea11.8%0%

La Roche-Posay is named in 100% of the answers of both AIs, but what follows diverges. ChatGPT puts Eucerin first in 53% of its answers, ahead of La Roche-Posay (41%), while Bioderma (82% presence) is never named first. Gemini puts La Roche-Posay first in 65% of cases, lifts SVR (82%) and Isdin (71%, against 18% on ChatGPT), and names a retailer in 100% of its answers, against 6% for ChatGPT.On the face, being named is not enough: Bioderma and Isdin are very present without ever being placed first.

What it means for a brand. The dermo-cosmetic core is already Visible and Mentioned. The challenge is the first mention, a signal of the Recommended level, won on usage promises (invisible under make-up, no white marks on dark skin) and on consistent product pages across the brand site and the retailers Gemini names in every answer.

Ranking: Body and family sun care

Base: 18 product-choice or problem prompts ranked, 12 advice prompts analysed separately.

How users search. These are parents and grandparents in a hurry: “my kids won't stay still on the beach, spray or milk, which is quickest?”, “my child refuses sun cream”. They compare formats (cream, spray, stick), shop for a baby under one or for atopic skin, and run into concrete irritants: sticky sand, stinging eyes.

Bar and first figure: presence rate (Mentioned). Light part of the bar and greyed figure: first mention (recommendation signal).

ChatGPT
  1. 1La Roche-Posay72.2%22.2%
  2. 2Bioderma61.1%27.8%
  3. 3Avène61.1%5.6%
  4. 4Garnier16.7%11.1%
  5. 5Mustela16.7%11.1%
  6. 6Alphanova11.1%0%
  7. 7Nivea5.6%0%
  8. 8SVR5.6%0%
  9. 9Uriage5.6%0%
Gemini
  1. 1La Roche-Posay100%44.4%
  2. 2Mustela72.2%5.6%
  3. 3SVR66.7%5.6%
  4. 4Avène61.1%0%
  5. 5Garnier44.4%5.6%
  6. 6Décathlon38.9%5.6%
  7. 7Bioderma38.9%0%
  8. 8Laboratoires de Biarritz33.3%5.6%
  9. 9Nivea33.3%0%
  10. 10A-Derma16.7%5.6%

Mustela is named in 72% of Gemini's answers but only 17% of ChatGPT's: the “baby” credentials exist for one and not for the other. ChatGPT relies on pharmacy brands (La Roche-Posay 72%, Avène and Bioderma 61%) and names no brand at all in 22% of its answers. Gemini puts La Roche-Posay first in 44% of cases and names a retailer in 78% of its answers, mostly mass retail (E.Leclerc, Carrefour, Intermarché, Auchan).For families, one ChatGPT answer in five contains no brand at all: the space is there to be taken.

What it means for a brand. Credentials for children are built situation by situation (age, format, beach, atopic skin) and have to exist in the sources ChatGPT reads, not only in Gemini's. Everyday irritants (stinging eyes, sand) are questions on which a brand can become the reference answer.

Ranking: Mineral and natural sun care

Base: 16 product-choice or problem prompts ranked, 14 advice prompts analysed separately.

How users search. This is the universe of doubt and verification: “how to spot nanoparticles in an INCI list”, “are reef-friendly creams really better for reefs?”. Add eczema, allergy to chemical filters, sport (surfing, hiking) and white marks. Nearly one prompt in two is an advice question.

Bar and first figure: presence rate (Mentioned). Light part of the bar and greyed figure: first mention (recommendation signal).

ChatGPT
  1. 1Avène68.8%25%
  2. 2Bioderma62.5%25%
  3. 3Laboratoires de Biarritz43.8%25%
  4. 4Uriage37.5%6.2%
  5. 5Alphanova25%18.8%
  6. 6La Roche-Posay18.8%0%
  7. 7Acorelle12.5%0%
  8. 8SVR12.5%0%
  9. 9Biosolis6.2%0%
  10. 10Cattier6.2%0%
Gemini
  1. 1Laboratoires de Biarritz93.8%62.5%
  2. 2Avène56.2%6.2%
  3. 3La Roche-Posay43.8%6.2%
  4. 4Acorelle31.2%0%
  5. 5Bioderma31.2%0%
  6. 6Patyka31.2%0%
  7. 7SVR25%6.2%
  8. 8A-Derma18.8%12.5%
  9. 9EQ Love18.8%0%
  10. 10Mademoiselle Bio18.8%0%

It is the only universe where La Roche-Posay does not lead (19% on ChatGPT, 44% on Gemini). Laboratoires de Biarritz dominates Gemini (94% presence, named first in 62% of answers) but stays at 44% on ChatGPT, where Avène (69%) and Bioderma (63%) take the lead. Gemini also names natural brands (Acorelle, Patyka, 31%) and a retailer in 81% of its answers (Biocoop, Naturalia, Sephora, Nocibé).In mineral, the top spot is up for grabs on ChatGPT: three brands share the lead, at 25% each.

What it means for a brand. On verification questions (nanoparticles, ocean, filters), the brand that explains best becomes the source. Natural brands are Visible on Gemini but little Mentioned on ChatGPT, and a dermo brand can take the “reactive skin or eczema” slot with a mineral range.

Ranking: Pharmacy and dermatology

Base: 18 product-choice or problem prompts ranked, 12 advice prompts analysed separately.

How users search. The search starts from a prescriber or an indication: “my dermatologist told me to wear SPF 50 every day, which brand?”, melasma, pregnancy mask, polymorphic light eruption, a treatment that makes skin sun-sensitive, a teenager under treatment. The user has already received advice and looks for the product that fits it.

Bar and first figure: presence rate (Mentioned). Light part of the bar and greyed figure: first mention (recommendation signal).

ChatGPT
  1. 1La Roche-Posay88.9%33.3%
  2. 2Bioderma83.3%27.8%
  3. 3Avène72.2%5.6%
  4. 4Eucerin66.7%22.2%
  5. 5SVR33.3%0%
  6. 6Isdin5.6%5.6%
  7. 7Uriage5.6%0%
  8. 8Vichy5.6%0%
Gemini
  1. 1La Roche-Posay100%44.4%
  2. 2Avène72.2%16.7%
  3. 3Bioderma72.2%5.6%
  4. 4SVR66.7%11.1%
  5. 5Eucerin38.9%16.7%
  6. 6Isdin33.3%0%
  7. 7Uriage22.2%0%
  8. 8Caudalie11.1%0%
  9. 9CeraVe11.1%0%
  10. 10Ducray11.1%0%

The most locked universe: La Roche-Posay (89% on ChatGPT, 100% on Gemini), Bioderma, Avène and Eucerin share the answers. Eucerin weighs 67% on ChatGPT against 39% on Gemini, SVR the reverse (33% against 67%), and Isdin rises to 33% on Gemini against 6% on ChatGPT. La Roche-Posay is named first in 33% of ChatGPT's answers and 44% of Gemini's.In pharmacy, the battle is for the first place, not for presence: five brands already share most of the answers.

What it means for a brand. A page per indication (melasma, light eruption, photosensitisation, teenager under treatment), validated by a dermatologist and backed by accessible clinical data, is what moves a brand from “named” to “named first”.

Ranking: After-sun and hair

Base: 17 product-choice or problem prompts ranked, 6 advice prompts analysed separately.

How users search. This is the most problem-driven universe, in the most spoken language: “my hair is like straw after the sea, what can I do”, “I have little itchy spots after the sun, what do I put on”. It includes the evening sunburn, peeling skin, the red child after the beach, hair bleached by the pool and curls to protect.

Bar and first figure: presence rate (Mentioned). Light part of the bar and greyed figure: first mention (recommendation signal).

ChatGPT
  1. 1Avène41.2%11.8%
  2. 2René Furterer29.4%11.8%
  3. 3Bioderma29.4%0%
  4. 4Klorane23.5%17.6%
  5. 5La Roche-Posay23.5%11.8%
  6. 6Phyto23.5%0%
  7. 7Garnier17.6%5.9%
  8. 8L'Oréal Paris17.6%0%
  9. 9CeraVe11.8%11.8%
  10. 10Kérastase11.8%5.9%
Gemini
  1. 1Avène58.8%23.5%
  2. 2La Roche-Posay58.8%17.6%
  3. 3Aroma-Zone41.2%11.8%
  4. 4Bioderma41.2%5.9%
  5. 5SVR41.2%0%
  6. 6Garnier35.3%5.9%
  7. 7Kérastase35.3%5.9%
  8. 8Nuxe35.3%5.9%
  9. 9Klorane35.3%0%
  10. 10Uriage35.3%0%

The most fragmented: no brand exceeds 59% presence or 24% first places. ChatGPT names Avène (41%), René Furterer (29%) and Bioderma (29%) and names no brand at all in 18% of its answers. Gemini blends skin and hair (Avène and La Roche-Posay 59%, Aroma-Zone 41%, Kérastase, Nuxe and Klorane 35%) and names a retailer in 59% of its answers.On after-sun, no brand dominates: the territory is open to whoever answers the problems in the users' own words.

What it means for a brand. It is the most open universe. Skin brands do not talk hair, hair brands do not talk after-sun: the one that owns the problem (“repair after the sea”, “soothe tonight's sunburn”) has a real chance of taking the first place.

Presence is not preference

The presence rate measures the Mentioned level of the V.M.R. methodology. First mention, the share of answers in which the brand is the first one named, is a recommendation signal: stronger than simple presence, it does not mean the AI explicitly claims to recommend the brand. Across the 86 ranked prompts, the gap between the two is the real measure of the work to be done.

BrandChatGPT presenceChatGPT first mentionGemini presenceGemini first mention
Bioderma64%16%48%2%
La Roche-Posay62%22%81%36%
Avène62%10%59%9%
Eucerin30%15%20%7%
SVR20%0%57%7%
Laboratoires de Biarritz8%5%28%13%
Mustela5%2%23%1%
Isdin5%1%23%1%

Bioderma is the most present brand on ChatGPT (64%) but is named first in only 16% of answers, and just 2% on Gemini. SVR is named in 57% of Gemini's answers, first in 7%, and never first on ChatGPT. La Roche-Posay is the brand most often named first on both AIs.

Advice or purchase: two AIs, two logics

On shopping prompts, ChatGPT nearly always names brands and never names a retailer, whereas Gemini names a retailer in nearly three answers out of four. On the 57 advice questions (“why does my sun cream pill under make-up?”, “how to spot nanoparticles?”), ChatGPT answers with no brand at all in 86% of cases while Gemini nearly always names some, with a retailer in almost one answer in two. On ChatGPT, brands only appear at the moment of choice. On Gemini, they are also played out in advice and on retailers' product pages.

Share of answersShopping (86 prompts) ChatGPTShopping GeminiAdvice (57 prompts) ChatGPTAdvice Gemini
No brand at all9%0%86%2%
At least one retailer2%73%2%46%

The 7 review questions, left out of the ranking

Seven prompts asked for an opinion on a specific product (Anthelios UVMune 400, Bioderma Photoderm, Vichy Capital Soleil, Supergoop, Typology, Uriage Bariésun, Sephora Collection). Gemini names a retailer in 100% of these answers and puts La Roche-Posay forward (86%). ChatGPT compares with competing brands (Avène and Bioderma 71%, La Roche-Posay 57%). When users ask whether a product is “worth it”, retailers are therefore a channel not to be neglected.

Key takeaway. In sun care, being named is not enough: Bioderma is the most present brand on ChatGPT (64%) but is named first in only 16% of answers, and La Roche-Posay is the brand most often named first on both AIs (22% and 36%). ChatGPT answers like an adviser and leaves advice questions brandless, Gemini answers like a shopping engine and leans on retailers: the two AIs have to be worked differently.

What the Radar measures, what the diagnostic explains

The Radar shows which brands AIs mention and put forward, at sector level. The AI Brand Echo diagnostic explains why, on your brand's own queries, and MYA continues the same method all the way to activation.

Public barometer, at sector level

AI Beauty Radar

  • MentionedPresence rate: share of answers in which the brand appears.
  • Recommendation signalFirst mention: share of answers in which the brand is the first one named.
  • What it does not showThe sources, content and signals behind these results.

Audit built on your brand's queries

AI Brand Echo diagnostic

  • VisibleThe content, pages, media, reviews and signals that make the brand known and understood by AIs.
  • MentionedThe queries on which the brand really enters the conversation.
  • RecommendedThe queries on which the brand becomes a preferred choice.

MYA's three-step method

Step 1

Diagnose

V.M.R. score on your key queries, AI by AI, with what the public Radar does not show: sources used, semantic territory and competitor benchmark.

Step 2

Prioritise

Building a panel of priority queries tailored to your brand, its uses and its competitors, spotting vacant territories and gaps between AIs, content roadmap.

Step 3

Activate

Recommendation pages by profile and use, structured FAQs, structured data and llms.txt, AI-friendly press content, reviews and forums, V.M.R. tracking over time.

What the Radar reveals, what MYA activates for your brand

  • Users searching for a feeling or a problem, not a brandRecommendation pages by profile and use, structured FAQs
  • ChatGPT leaving advice without brandsExpert content and AI-friendly press content, op-eds and experts
  • Gemini leaning on retailersConsistent retailer product pages, reviews, structured data
  • Marked gaps between the two AIsV.M.R. tracking AI by AI, over time

The Radar shows what is happening. AI Brand Echo explains why. MYA acts to change it.

See the three-step method

What about your brand?

The Radar shows which brands AIs put forward. The AI Brand Echo diagnostic explains why, on your brand's own queries.

Request my V.M.R. diagnostic
← Issue #1Issue #3 →