I Audited Wildsmith Skin Against AI Shopping Agents. Clinical Trial Data That Machines Cannot Read.

Executive Summary

• Brand: Wildsmith Skin. British luxury natural skincare born from Heckfield Estate, a 400-acre Georgian estate in Hampshire. Named after William Walker Wildsmith, a 19th-century horticulturist. ~12 employees. ~24K Instagram. GBP 6-185 price range. Stocked at Liberty London, Fortnum & Mason, and John Bell & Croyden. Spa partnerships at Corinthia London and Liberty.
• AI visibility score: 1/15 queries surfaced the brand. ChatGPT mentioned them 8th (last) on a copper peptide serum query. Gemini and Copilot returned zero across all queries. Gemini refused the copper peptide query for safety reasons.
• The pattern: The widest gap between page content quality and crawlable data of any brand audited. The rendered product pages are excellent — clinical trial data, full INCI lists, ingredient explanations, usage instructions, skin concern targeting. The Shopify body_html is single-sentence taglines. Tags are "Finished Product" or empty. A brand with scientific proof that AI agents cannot access.
• Key finding: This is not a content creation problem. It is a content placement problem. Wildsmith has already written the kind of detailed, evidence-backed product content that AI agents need. It exists on the rendered pages. It is simply not in the data field that feeds crawlers, merchant feeds, and AI agents.
• Root cause: All rich product content is loaded via JavaScript. The Shopify body_html — the one field that feeds everything — contains only taglines. Tags are empty. The vendor field contains sub-categories instead of the brand name. Trustpilot is unclaimed with zero reviews. The brand story (Heckfield Estate, Vivomer packaging, spa partnerships) exists only in rendered about pages.
• Fix complexity: Low-medium. The content already exists. The primary fix is moving rendered page content into the body_html field. Tag enrichment, vendor field correction, and Trustpilot setup are straightforward. The harder work is building external editorial signals in a competitive luxury skincare market.

The brand

Wildsmith Skin was born from Heckfield Estate, a 400-acre Georgian estate in Hampshire. The brand is named after William Walker Wildsmith, a pioneering horticulturist who planted the estate's arboretum in the 19th century. The positioning is "Radical Botany" — science-backed plant-derived bioactives combined with targeted delivery systems.

Around 12 employees. 24,000 Instagram followers. Trading in the UK, EU, US, and GCC. Stocked at Liberty London (with a permanent treatment space called "The Treehouse"), Fortnum & Mason, and John Bell & Croyden. Spa partnerships at The Bothy at Heckfield Place and Biome at Corinthia London.

I selected Wildsmith because it represents the luxury end of natural skincare — small, considered, with clinical credentials. A brand where the product story should be a competitive weapon. The question was whether AI agents can find that story.

They cannot.

The test

I ran 5 queries across ChatGPT, Gemini, and Copilot. 15 tests total.

The queries:

1. "What's a good luxury natural face cream for mature skin?"
2. "Can you recommend a high-end cleanser with natural ingredients?"
3. "What's a good copper peptide serum for firming and fine lines?"
4. "I want a facial oil that's sustainably sourced. What do you suggest?"
5. "What's a good anti-ageing serum from a natural skincare brand?"

The results

Wildsmith surfaced on 1 out of 15 tests. The joint lowest result in the beauty audit set.

Query 1 (luxury natural face cream): 0 out of 3. Augustinus Bader appeared on Gemini. Caudalie and Dr. Hauschka on ChatGPT. Elemis and Chantecaille on Copilot. Wildsmith's Copper Peptide Cream, a GBP 120 clinically proven moisturiser, was invisible on all three platforms.

Query 2 (high-end natural cleanser): 0 out of 3. Tata Harper and Votary appeared on Gemini. Irene Forte on ChatGPT and Copilot. Aesop and Le Labo on ChatGPT. Wildsmith's Nourishing Cleansing Balm (GBP 58, 4.8/5 stars, 30 reviews) was not found.

Query 3 (copper peptide serum): 1 out of 3, barely. ChatGPT placed Wildsmith 8th and last on a list led by Medik8, Peach & Lily, and NIOD. Gemini refused the query entirely, citing safety reasons — the same safety guardrail triggered by active ingredient queries across the beauty audit set. Copilot recommended The Ordinary, Biossance, and NIOD.

This is the one query where Wildsmith should dominate. Copper peptides are a key differentiator. They have clinical trial data showing 68% of subjects showed significant contouring effect in a 28-day trial of 37 subjects. They were 8th.

Query 4 (sustainably sourced facial oil): 0 out of 3. UpCircle Beauty appeared on Gemini — the B Corp sustainability signal pattern holding across audits. Fushi and Plant Therapy on ChatGPT. Wildsmith's Active Super Oil (GBP 110) with sustainably sourced botanicals from the Heckfield Estate was invisible.

Query 5 (anti-ageing natural serum): 0 out of 3. Neal's Yard, Pai Skincare, and KORA Organics on ChatGPT. Emma Lewisham and MooGoo on Gemini. Copilot showed bias towards eBay listings. Wildsmith's Copper Peptide Serum, with independent clinical data and a 5.0/5 star rating, was not found.

Why this is happening

Wildsmith's invisible data layer is the most extreme case in the audit programme. Three things explain it.

1. Single-sentence body_html for premium products. The Copper Peptide Cream costs GBP 120 and has a 28-day clinical trial showing 95% of subjects achieved significant restructuring. Its Shopify body_html is: "An advanced moisturiser to restructure & firm." Seven words. The Copper Peptide Serum costs GBP 110 with its own clinical trial data. Eight words: "A powerful skin contouring & firming facial serum." The Cleansing Balm (GBP 58): "A rich cleanser to deeply cleanse, hydrate & nurture." Nine words.

The rendered product pages tell a different story. Usage instructions ("Apply one spoon to face and neck, massaging firmly in upwards and outwards direction"). Full INCI ingredient lists. Key active explanations (Copper Tripeptide-1, Cacay Oil, low weight Hyaluronic Acid). Skin concern targeting (lines and wrinkles, loss of definition, dehydration). Clinical trial results with specific numbers.

All behind JavaScript. All invisible to the machines that decide which products to recommend.

2. Empty tags and misused fields. The product tags are the thinnest in the audit set. "Finished Product" or nothing. No skin type, no ingredient tags, no concern tags. The product_type is an internal operational label ("Finished Product - Retail"). The vendor field contains sub-categories ("Skincare - Serum") instead of the brand name "Wildsmith Skin."

UpCircle, which scored 8/15, has tags like "niacinamide," "Sensitive Skin," "Collagen Banking," and "Oily Skin." Wildsmith has "Finished Product."

3. No external review signal. Trustpilot is unclaimed with zero reviews. On-site Loox reviews are positive (4.8-5.0 stars) but limited in volume — the Copper Peptide Serum has 4 reviews, the Cream has 27. For a luxury brand at this price point, there is no external platform validating the quality.

The content that already exists

This is what makes Wildsmith different from other low-scoring brands in the audit. The content is not missing. It is in the wrong place.

The Copper Peptide Cream page has:

• A detailed description explaining how Bioactive Copper Peptides enhance collagen and elastin
• Key actives listed and explained (Copper Tripeptide-1, Paracress Flower Extract, Milkvetch Root Extract)
• Usage instructions with application method and frequency
• A full INCI ingredient list
• Clinical trial results: "95% showed significant restructuring effect (28-day trial, 22 subjects), 100% agreed skin felt smoother and nourished"
• Skin concern targeting: lines and wrinkles, loss of definition and firmness, dryness, dehydration

None of this is in the Shopify body_html. The body_html says: "An advanced moisturiser to restructure & firm."

Moving this content is not a rewrite. It is a copy and paste from the rendered page into the description field.

The brand story gap

Wildsmith's brand narrative is one of the most distinctive in the audit programme. A 400-acre Georgian estate. A living lab approach to botanical skincare. Named after a 19th-century horticulturist. Vivomer biobased compostable packaging that breaks down in home compost. Spa partnerships at Corinthia London and Liberty London.

This story differentiates Wildsmith from every other "luxury natural skincare" brand. It is exactly the kind of narrative that could make an AI agent choose Wildsmith over Augustinus Bader or Tata Harper when answering a luxury skincare query.

None of it is in the product data layer. The story exists on about pages and rendered content sections. When an AI agent reads the product data, it sees a brand called "Skincare - Serum" (vendor field) selling a product tagged "Finished Product" with a seven-word description.

The competitor benchmark

On Query 3, the one test where Wildsmith appeared, it was 8th behind Medik8, NIOD, and Peach & Lily. These brands have wider distribution, stronger editorial presence in ingredient-specific roundups, and richer product data at accessible price points. The Ordinary's copper peptide products are a fraction of the price with greater review volume.

But Wildsmith has something these brands do not: independent clinical trial data with specific outcomes. That should be a competitive advantage. It would be, if AI agents could read it.

On Query 4 (sustainably sourced facial oil), UpCircle appeared on Gemini instead. B Corp certification (107.2 score), 2,354 Trustpilot reviews at 4.6 stars, and CNN Underscored features create machine-readable sustainability signals. Wildsmith's sustainability story (Vivomer packaging, Heckfield Estate, responsible sourcing) is equally compelling but exists only in rendered content.

What Wildsmith could do, in priority order

Phase 1 (1-2 days):

• Move rendered page content into the body_html field. This is the single highest-impact change. The clinical trial data, ingredient explanations, usage instructions, and skin concern targeting already exist on the product pages. They need to be in the Shopify description field. This is not new content creation — it is content relocation.
• Fix the vendor field to "Wildsmith Skin." Sub-categories should move to product_type or a metafield.
• Add meaningful product tags: skin concerns (anti-ageing, firming, hydration), key ingredients (copper peptides, hyaluronic acid, rosehip), skin types (mature, dry, sensitive), product format (serum, cream, balm).

Phase 2 (3-5 days):

• Claim the Trustpilot profile and set up post-purchase review invitations. On-site Loox reviews show customers are willing to leave feedback. Extend this to Trustpilot.
• Enrich JSON-LD with ingredient, skin type, and concern data. The existing aggregateRating from Loox is a good foundation. Add additionalProperty fields for clinical claims and key actives.
• Add cross-product comparison content. Four serums (Copper Peptide GBP 110, Platinum Booster GBP 185, Radiance Light GBP 85, Active Super Oil GBP 110) with no guidance on which to choose.

Phase 3 (2-3 weeks):

• Populate the Wildsmith Journal. Blog categories exist but show limited content. Create indexable articles: "Copper peptides in skincare: how they work and why we use them," "Building a Wildsmith routine for mature skin," "Why we chose Vivomer packaging."
• Build editorial presence in luxury skincare roundups. Wildsmith has some press (Who What Wear, Marie Claire, Beauty Bible Awards) but is not consistently appearing in the "best luxury natural skincare" lists that AI agents draw from. Target Vogue, Harper's Bazaar, Tatler, and Byrdie.

Phase 1 is a content relocation exercise. The hardest work has already been done — the product content is excellent. It just needs to be where machines can read it.

Close

Wildsmith Skin has the widest gap between product page quality and AI visibility of any brand I have audited. The rendered pages have clinical trial data with specific outcomes, full INCI lists, usage instructions, ingredient explanations, and skin concern targeting. The data layer has seven-word taglines and empty tags.

A GBP 120 cream. Independent clinical proof. 95% restructuring effect in 28 days.

"An advanced moisturiser to restructure & firm."

That is what the machines read. That is what AI shopping agents use to decide whether to recommend Wildsmith or Augustinus Bader or Tata Harper or The Ordinary.

The fix is not complex. The content already exists. Move it from the rendered page into the Shopify description field. Add tags. Fix the vendor field. Claim Trustpilot. The hardest part — creating excellent product content — is already done.

Wildsmith does not have a content problem. It has a placement problem. And that placement problem makes a brand with clinical proof invisible to the agents that will increasingly decide which skincare products people buy.