Research

AI Advent Calendar Index 2026: the brands AI recommends to UK shoppers

More than 330,000 AI visibility data points reveal how advent calendar brands and retailers are mentioned, recommended and described in AI-generated Christmas shopping answers.

Josh Tulip, Founder, Citations.ioBy Josh Tulip · Founder, Citations.io

Research from Citations.io's AI Advent Calendar Index 2026, examining more than 330,000 recorded AI visibility data points across four engine-labelled samples, reveals how AI engines represented and recommended brands and retailers in sampled UK shopping questions for Christmas 2026.

The brands leading AI recommendations this Christmas

Different shopping questions produce different leaders. The following findings use selected 2026-specific unbranded shopping questions from the OpenAI-model sample.

  • General Christmas and gift calendars: Fortnum & Mason — 43.8% recommendation rate, 43.8% visibility, 100% positive classified sentiment.
  • Beauty: Boots — 37.5% recommendation rate, 59.0% visibility, 55.1% positive sentiment.
  • Budget beauty: Superdrug — 72.5% recommendation rate, 95.0% visibility, 68.4% positive sentiment.
  • Premium beauty: Selfridges — 36.0% recommendation rate, 69.0% visibility, 46.4% positive sentiment.
  • Children's calendars: LEGO — 68.4% recommendation rate, 83.5% visibility, 81.8% positive sentiment.
  • Chocolate: Hotel Chocolat — 73.1% recommendation rate, 73.1% visibility, 100% positive sentiment.
  • Jewellery: Monica Vinader — 65.5% recommendation rate, 65.5% visibility, 94.4% positive sentiment.
  • Home and fragrance: Jo Malone London and Diptyque — both 48.4% recommendation rate, 50.0% visibility and 96.9% positive sentiment.

Visibility is not the same as recommendation

John Lewis achieved 72.0% visibility across the pooled category sample, but a recommendation rate of 37.9%. Brands can appear in answers for different reasons, including as retailers, comparators or suggested products. Tracking visibility alone does not tell the whole story.

How positively does AI discuss Christmas shopping brands?

Across the full study, 64.0% of classified brand mentions were positive, 35.9% neutral and 0.1% negative. The overwhelmingly non-negative tone reflects a shopping recommendation task and the automated sentiment extraction process; it should not be interpreted as consumer satisfaction.

Explore the category research

What the index reveals

Visibility, recommendation strength, sentiment and position are distinct measures. The results show the value of examining the context in which a brand appears, not just its presence in a generated answer. The study does not rank individual calendar editions, establish current stock or assess product quality.

Methodology and limitations

Study scale: 334,830 recorded data points comprising 79,541 brand mention records, 79,541 recommendation classifications, 79,532 sentiment classifications, 39,777 non-null answer position measurements and 56,439 extracted source reference records. URL and domain representations of a source are not counted separately. These data points are measured attributes of sampled AI responses from a single point-in-time collection; they are not independent observations or an answer count. Source references are extracted and are not necessarily verified or native citations.

Analysis: The findings use saved neutral-market prompts explicitly referencing 2026, excluding purchase-location and availability intent. The category findings use relevant subsets of the ChatGPT-labelled group, sampled with OpenAI GPT-5 mini via API, not the consumer ChatGPT interface. AI visibility is the percentage of relevant answers containing a brand mention; recommendation rate is the percentage containing a brand flagged as recommended. Both use distinct answers and stored canonical brand labels, excluding flagged off-topic records. Positive, neutral and negative sentiment are percentages of classified brand mentions; unknown labels are excluded. Sentiment is automated interpretation of generated text, not human consumer sentiment, and may correlate with the recommendation classifier. Multiple brands can appear per answer, and category subsets overlap. The study is not weighted to actual shopper behaviour.

The figures describe a preliminary historical sample; extraction and alias matching can introduce errors. The study does not provide live rankings or verify current product availability, price, quality or sales. The overall data point count describes the full study, not the denominator of each headline. Repeated answers to the same question are not independent; close rankings and smaller subsets are directional. Exact sample sizes and category denominators are available on request for editorial or research verification.

Media enquiries

Citations.io is an AI citation intelligence platform helping brands understand how AI engines mention, rank and cite them. For research enquiries, supporting data or commentary, contact Josh Tulip, founder, at support@citations.io.

ResearchAI Advent Calendar Index 2026AI visibilityChristmas shopping

Related briefs

Supported by
Google for StartupsNatWest Accelerator