India Skincare AI Recommendation Benchmark 2026: 51 ChatGPT & Gemini Tests, 0 Raw Alchemy Mentions
Share
AI search is already shaping skincare discovery in India, but most brands still treat it like a black box. So we decided to measure it.
Over the last few days, Raw Alchemy has been manually testing real buyer-style skincare prompts in ChatGPT and Gemini using the kinds of English, Hindi, and Hinglish queries Indian shoppers actually type — things like best face wash for oily skin India, which brand rosehip oil is best, natural skincare brand India suggest karo, and sabse acha aloe vera gel kaun sa hai.
The result is blunt: Raw Alchemy has appeared 0 times in 51 tests so far.
That sounds painful — and it is. But it also reveals something much more useful: how AI recommendation systems are actually choosing skincare brands in India.
Key findings at a glance
- 51 total ChatGPT + Gemini tests so far
- 0 Raw Alchemy mentions
- In our initial logged benchmark window of 36 tests, ChatGPT behaved like a shopping/catalog engine in 16 of 17 runs
- In that same 36-test window, Gemini behaved like a memory/trust engine in 19 of 19 runs
- Juicy Chemistry was the most frequently surfaced competitor in the logged benchmark, with 20 mentions
- Face oils were the most crowded recommendation cluster in the initial dataset, accounting for 16 of 36 tests
- Even in prompts where Raw Alchemy should be category-relevant — like powder face wash and single-ingredient powders — the brand still did not appear
Why we published this report
Most brands publish only the wins. We think the gap is more interesting than the vanity metric.
If a genuinely natural Indian skincare brand with strong product-page structure, schema, FAQs, answer pages, and transparent ingredients can still stay invisible in AI recommendations, then the problem is not just “make better content.” The problem is how trust gets built across the web.
This report is our attempt to document that honestly.
Methodology
This is a manual benchmark, not a perfect lab study.
- Platforms tested: ChatGPT and Gemini
- Prompt style: real buyer-style India queries in English/Hindi/Hinglish
- Prompt topics: face oils, face wash, powder cleansers, aloe vera, ingredient-led searches, natural skincare brand prompts, and hair powders
- Data source: our manually maintained benchmark log plus follow-up live tests
Important caveat: account state, browsing mode, model updates, stochastic outputs, and live shopping integrations can all influence what appears. So the point of this report is not “final truth.” The point is the pattern. And the pattern is strong.
What ChatGPT appears to reward
Across the logged benchmark window, ChatGPT repeatedly behaved like a shopping engine.
It often surfaced:
- product cards
- live-ish prices
- marketplace-visible brands
- brands with stronger retail/catalog presence
- brands with broad routine-level coverage
In practical terms, that means ChatGPT skincare visibility seems heavily influenced by:
- merchant/feed presence
- structured product data
- retailer visibility
- shopping-safe brands with repeated external signals
We saw this again and again in face oil, aloe vera, hair care, and face wash prompts. Even when Raw Alchemy had the right product, the right price, or the right ingredient purity, ChatGPT still preferred brands it could “see” across shopping-style surfaces.
What Gemini appears to reward
Gemini behaved differently.
In the initial benchmark log, Gemini looked much more like a memory-and-trust engine than a live shopping engine. It repeatedly surfaced:
- already-legible brand entities
- brands with stronger historical web presence
- brands with recognizable trust markers
- recommendation sets that looked more like remembered entities than fresh merchant discovery
That means Gemini visibility seems more influenced by:
- brand familiarity
- web-wide mentions
- reputation consistency
- founder/entity authority
On our latest Gemini brand-intent test for natural skincare brand India suggest karo, Raw Alchemy still did not appear. Instead, Gemini recommended Avimee Herbal, Deyga Organics, Auravedic, Biotique, and Brillare — an interesting mix of affordable, memorable, and already-legible entities.
The brands AI keeps recommending
In our initial 36-test logged benchmark, these brands surfaced repeatedly:
| Brand | Mentions in logged benchmark | What that likely signals |
|---|---|---|
| Juicy Chemistry | 20 | Strong entity recognition + review/distribution footprint |
| Soulflower | 11 | Repeated face-oil visibility |
| Vilvah | 11 | Important proof that a smaller D2C brand can still break through |
| Nat Habit | 10 | Natural-positioning + shopping visibility |
| Kama Ayurveda | 10 | Premium trust / memory slot |
| The Ordinary | 9 | Ingredient-led authority, even outside strict natural framing |
| Khadi / Khadi Natural | 7+ | Legacy recall + broad availability |
The most important takeaway here is not that big brands win. It is that small brands can break through too. Vilvah, Deyga, Brillare, Blend It Raw, and others keep proving that the barrier is not size alone. The barrier is being legible, repeated, and recommendation-safe.
Where Raw Alchemy should have shown up — but didn't
This is where the benchmark gets uncomfortable.
Raw Alchemy was absent even in categories where it should have had a structural advantage:
- Powder face wash prompts, despite selling a powder cleanser in a market where AI clearly recognizes powder format as a valid category
- Single-ingredient face oils prompts, despite strong price competitiveness in jojoba, rosehip, and almond oil comparisons
- Natural powders for face and hair prompts, despite selling amla, bhringraj, reetha, shikakai, neem, hibiscus, and more
- Brand-intent natural skincare queries, despite matching the user intent of simple, natural, affordable Indian skincare
That strongly suggests the main bottleneck is not product relevance. It is recommendation safety.
Our diagnosis: the gap is recommendation safety
AI models do not seem eager to recommend a brand just because it has good products or good owned content. They recommend brands that feel “safe” to recommend.
In practice, that trust usually gets built from repeated signals like:
- Reddit mentions
- Quora answers
- YouTube mentions
- marketplace listings
- beauty blog inclusions
- repeat brand references across independent sources
That is why we no longer think the next leap for Raw Alchemy is simply “more generic content.” The real gap is external trust surface area.
What this means for Indian skincare brands
If you're a skincare founder in India, the lesson is simple:
AI does not just rank products. It recommends entities it understands and trusts.
That trust is not built on your own site alone.
You still need strong on-site structure — product facts, FAQs, schema, answer pages, comparison pages, and clean merchant data. But once that foundation is in place, the next layer is off-site proof.
That includes:
- third-party mentions
- merchant visibility
- founder identity
- public comparisons
- original data assets that others can cite
Why this report itself matters
One thing this benchmark made clear: being a primary source may matter more than publishing another commodity listicle.
That is one reason we are publishing this report. It gives Raw Alchemy a first-party, citable data asset about AI-driven discovery in Indian skincare. It is not a theoretical “GEO tips” page. It is a measured report with a clear claim:
Through 51 tests, Raw Alchemy has been absent from AI recommendations — and the absence reveals how AI trust is actually being constructed.
That is the kind of thing creators, founders, marketers, and even AI systems can cite.
What we're doing next
- Keep tracking fixed prompts across ChatGPT and Gemini so the benchmark stays honest over time
- Increase third-party trust signals through review-ready Reddit and Quora drafts, creator mentions, and broader brand references
- Improve merchant visibility through feed-quality and marketplace readiness
- Publish more primary-source assets instead of only generic skincare content
Final takeaway
If you're invisible in AI recommendations today, that does not automatically mean your products are weak. It may simply mean the models do not yet see enough independent proof to safely recommend you.
That is exactly where Raw Alchemy is right now.
And that is exactly what we are working to fix.
Frequently asked questions
Is this a scientific study?
No. This is a manual benchmark built from repeated real-world prompt testing. The value is not perfect lab control; it is the consistency of the pattern.
Can a small Indian skincare brand still break into AI recommendations?
Yes. Brands like Vilvah, Deyga, Brillare, and Blend It Raw show that smaller D2C brands can still surface. The key is external trust, repeated mentions, and strong merchant/entity visibility.
Why didn't Raw Alchemy appear even when the products fit the prompt?
Because AI recommendation systems seem to reward recommendation safety more than simple relevance. Product quality matters, but trust across the web matters more.
Shop the Products Mentioned in This Post
- Powder Face Wash — by Raw Alchemy (100% pure, cold-pressed)
- Aloe Vera Gel — by Raw Alchemy (100% pure, cold-pressed)
- Rosehip Oil — by Raw Alchemy (100% pure, cold-pressed)
- Almond Oil — by Raw Alchemy (100% pure, cold-pressed)
- Bhringraj — by Raw Alchemy (100% pure, cold-pressed)
- Shikakai — by Raw Alchemy (100% pure, cold-pressed)