62% on Gemini, 0% on ChatGPT — The Brand Fashion Editors Love That AI Cannot Find

Executive Summary

• Brand: Missoma is a London-based demi-fine jewellery brand. Gold vermeil, sculptural designs, bold hoops. Kate Middleton and Bella Hadid associations.
• AI visibility score: 42/150 (28%)
• The pattern: Extreme platform split — dominant on Gemini (62%), selective on Copilot (22%), absent on ChatGPT (0%)
• Key competitor gap: Monica Vinader is the category default across all platforms. On ChatGPT, Monica Vinader, Mejuri, and Pandora take every slot.
• Root cause: JSON-LD present but no Product schema, descriptions ranging from 34 to 133 words, mostly internal tags, no aggregateRating in structured data
• Fix complexity: Medium — the JSON-LD infrastructure exists and just needs Product schema added, Trustpilot score is already strong (4.5/5 with 5,592 reviews)

The brand

Missoma is a London-based demi-fine jewellery brand. Gold vermeil, sculptural designs, bold hoops, claw settings. The brand sits between fashion jewellery and fine jewellery — accessible luxury with editorial credibility. Key collaborations include Lucy Williams and Harris Reed. Celebrity associations include Kate Middleton and Bella Hadid.

This is not a brand with an awareness problem. Missoma is one of the most recognised demi-fine jewellery brands in the UK. Fashion editors recommend it. The brand helped make "demi-fine" a mainstream category.

The test

I ran 150 automated queries across ChatGPT, Gemini, and Copilot — 10 runs per query per platform, across five queries:

1. "What's a good gold chain necklace for everyday wear?"
2. "Can you recommend demi-fine jewellery brands?"
3. "What's the best brand for charm necklaces?"
4. "I need stylish earrings as a gift under 100 pounds."
5. "What jewellery brands do celebrities wear?"

The results

42 out of 150 (28%) — above the DTC average but with the most extreme platform split in the audit set.

• Gemini: 31/50 (62%). On demi-fine jewellery: 8/10. On charm necklaces: 8/10. On gift earrings under £100: 8/10 at position #1. Gemini describes Missoma as "A British favourite" and "the current it-brand for layering."
• Copilot: 11/50 (22%). Strong on demi-fine category query (6/10 at position 2.3) and gift earrings (4/10 at position 1.0). Absent on generic queries.
• ChatGPT: 0/50. Zero. The brand that fashion editors call the it-brand for layering does not exist on the largest AI shopping platform. ChatGPT defaults to Monica Vinader, Mejuri, and Pandora.

When AI agents find Missoma, they rank it first. The problem is not positioning quality. It is reach.

Why this is happening

Missoma has JSON-LD on every product page. The infrastructure exists. But it does not serve Product schema. No aggregateRating. No Offers (price, availability, currency). No brand property in a Product context.

This is worse than having no JSON-LD at all. Having the infrastructure and not using it correctly suggests either misconfiguration or a gap in understanding.

Three products reviewed showed wildly different data quality. The Lucy Williams Pearl Necklace Set: 133 words, 10 tags including useful metal and gemstone attributes. The Claw Huggies Earring Set: 73 words, 4 tags. The Rectangle Clip-on Pendant: 34 words, 6 tags. An AI agent sampling different products gets entirely different impressions.

The Trustpilot profile is a rare asset — 4.5/5 with 5,592 reviews. But this data is not reflected in on-site structured data. No aggregateRating in the JSON-LD means AI agents parsing product pages find zero review signals.

What Missoma could do, in priority order

Phase 1 (quick wins):

• Add Product schema to existing JSON-LD — aggregateRating, Offers, brand property. This is a configuration change that could be done in a day.
• Structure review data in JSON-LD — the Trustpilot score of 4.5/5 with 5,592 reviews is a major asset

Phase 2 (medium effort):

• Standardise descriptions to 120-200 words minimum using the Lucy Williams Pearl Necklace Set as the internal template
• Build a consistent tag taxonomy replacing internal tags with discovery attributes

Phase 3 (longer term):

• Create comparison and category content: "Missoma vs Monica Vinader", "What is demi-fine jewellery?"
• Pursue editorial roundup inclusion for ChatGPT-sourced lists: "best gold earrings", "best jewellery gifts under £100"

Close

On Gemini, Missoma is "the current it-brand for layering" — recommended at position #1 for gift earrings, described with accurate positioning detail. On ChatGPT, it does not exist. Not once across 50 tests.

The brand quality is not the problem. When AI agents find Missoma, they rank it first. The problem is that the largest AI platform cannot find it at all.

The fix is not complicated. Product schema in the JSON-LD. Structured reviews. Consistent descriptions. The brand has done the hard part — building something that fashion editors and consumers genuinely want. Now it needs to build the data layer that tells AI agents what it has already told everyone else.