Who AI recommends — and why cross-cluster authority is rarer than it looks
Key Findings
Overall Findings
Gemini leads overall with 1,654 total mentions across all brands versus Claude (1,162), ChatGPT (1,141), and Perplexity (995) — a consistent pattern across identity-forward category audits. Platform concentration gaps point almost uniformly to Perplexity.
| Brand | ChatGPT | Claude | Gemini | Perplexity | Total |
|---|---|---|---|---|---|
| Ridge | 65 | 42 | 124 | 49 | 280 |
| McBride Sisters (consolidated) | 62 | 40 | 66 | 36 | 204 |
| Littorai | 11 | 20 | 55 | 15 | 101 |
| Turley Wine Cellars | 30 | 20 | 37 | 9 | 96 |
| Brown Estate | 34 | 10 | 24 | 17 | 85 |
| Bedrock | 12 | 23 | 35 | 6 | 76 |
| Maison Noir | 30 | 21 | 4 | 9 | 64 |
| Tablas Creek | 24 | 5 | 23 | 6 | 58 |
| Arnot-Roberts | 9 | 18 | 20 | 7 | 54 |
| Longevity Wines | 27 | 4 | 21 | 1 | 53 |
| Broc Cellars | 11 | 18 | 20 | 2 | 51 |
| Benziger | 9 | 17 | 14 | 10 | 50 |
| Bonterra | 19 | 9 | 12 | 9 | 49 |
| Quivira Vineyards | 15 | 15 | 7 | 11 | 48 |
| Seghesio | 10 | 12 | 11 | 15 | 48 |
Platform Concentration Gaps
Brands with zero Perplexity mentions despite meaningful presence elsewhere. This is not a coincidence of brand-level content decisions — it reflects a structural behavior of Perplexity's retrieval model, which indexes differently from transformer-based platforms.
| Brand | Other Platform Mentions | Gap Platform | Gap Mentions |
|---|---|---|---|
| Stag's Leap Wine Cellars | 42 | Perplexity | 0 |
| The Prisoner | 39 | Perplexity | 0 |
| Ravenswood | 27 | Perplexity | 0 |
| Scribe Winery | 26 | Perplexity | 0 |
| Apothic | 25 | Perplexity | 0 |
Cluster Analysis
This audit used five clusters specifically designed to surface identity-forward and craft-forward producers. The results reveal which content types drive cross-cluster authority — and why so few brands achieve it.
McBride Sisters' 204-mention total is the product of a deliberate multi-format brand strategy: SHE CAN, Black Girl Magic, and the core McBride Sisters line each carry the brand's identity into distinct consumer contexts, and AI systems index all of them. A defining characteristic of this cluster is its insularity — with the exception of Maison Noir, none of the top performers cross meaningfully into Sustainability, Terroir, or Occasion clusters. Visibility here is earned through identity-forward editorial content and explicit brand language around ownership and mission.
Ridge leads this cluster despite not being primarily positioned as a sustainability brand — a consequence of its deep content footprint across vineyard-specific farming practices and minimal-intervention winemaking philosophy. Littorai's showing reflects detailed site-specific biodynamic content. Benziger and Bonterra benefit from certification language — Demeter and CCOF respectively — that functions as a structured signal AI systems can classify reliably. Brands farming regeneratively without publishing farm-specific content are systematically absent here.
This cluster rewards brands that have published content explicitly connecting their wines to occasions, gatherings, and accessible enjoyment. Ridge surfaces through its long-standing populist positioning and the breadth of content connecting its wines to everyday drinking as well as special occasions. Notable for what's absent: identity-forward brands with strong anti-pretension brand language but minimal occasion-specific content are largely missing, suggesting that brand voice alone does not substitute for editorial content that explicitly names occasions and contexts.
Ridge's 124-mention lead is the largest single-brand total in the entire audit — reflecting decades of site-specific content: named vineyard designates, documented farming history, and winemaker writing that gives AI systems attributable, specific claims to index. This is the highest-volume cluster in the audit (1,117 total mentions), confirming that terroir and craft queries generate more AI recommendations than any other category tested. The content type driving those recommendations is specific, attributable, and place-grounded.
The lowest-volume cluster in the audit (801 total mentions) — reflecting the relative scarcity of structured tasting room and wine club content across California producers. Ridge and Jordan surface through a combination of well-documented visitor experience content and established wine club programs with clear membership descriptions. SommSelect and The California Wine Club are subscription-first platforms whose entire content architecture is organized around discovery and delivery — giving them a structural advantage here. The pattern is consistent: brands that publish explicit, structured content about their tasting room experience, reservation process, and club tiers earn inclusion. Brands with active programs but minimal on-site description of them do not.
What Drives AI Visibility in California Wine
Across 841 unique brand entities detected in this audit, two structural patterns dominate: a handful of cross-cluster authorities accumulate visibility across every query type, while the majority — including those with significant press coverage and distribution — remain confined to a single cluster or surface inconsistently across platforms.
Brands that name their vineyards, varietals, and farming practices in published content earn terroir cluster inclusion. The Terroir & Craft Discovery cluster is dominated by producers who have published content that gives AI systems specific, attributable claims to index: Ridge's Monte Bello and Lytton Springs vineyard histories, Turley's old-vine Zinfandel field documentation, Bedrock's heritage vineyard stewardship, Littorai's coast-specific site writing.
The specificity threshold matters. Generic language about "small-lot production" or "hand-harvested grapes" does not function as a recommendation signal. Named vineyard blocks, documented farming decisions, and varietal-specific tasting language do.
Brands that publish across multiple content formats and product expressions accumulate mentions at scale. Single-format brands plateau early regardless of editorial depth on any one dimension. The McBride Sisters' 204-mention consolidated total is the clearest illustration: SHE CAN, Black Girl Magic, and the core McBride Sisters line each create separate indexing surfaces — different products with distinct positioning, occasion language, and audience framing — that collectively give AI systems more entry points.
Ridge demonstrates the same principle across content formats rather than product lines: winery editorial, winemaker letters, vineyard site pages, retail listings, press coverage, and wine club content all create overlapping indexing surfaces that compound into cross-cluster dominance.
Perplexity performance is determined by citation density in indexed publications — and it is the most reliable leading indicator of a brand's authority-layer eligibility across all platforms. The five brands with Perplexity concentration gaps are not unknown brands — they have meaningful AI training data visibility confirmed by their ChatGPT and Claude scores. What they lack is the specific kind of third-party editorial coverage that Perplexity retrieves at query time.
The brands with the strongest Perplexity scores — Ridge (49), McBride Sisters (36), Brown Estate (17) — share consistent coverage in outlets that function as Perplexity citation sources. This pattern makes Perplexity performance a useful diagnostic: brands that score well have built authority-layer presence; brands that score zero have not, regardless of how well they perform on platforms that rely more heavily on training data alone.
No identity-forward Black-owned California wine brand currently operates with multi-cluster AI visibility. This is a content architecture gap, not a brand credibility gap. Producers with regenerative farming credentials, named vineyard sites, and accessible occasion-forward positioning have the raw material for cross-cluster authority — but have not yet published the content that would allow AI systems to classify them across categories simultaneously.
What California Wine Brands Can Do With This
A brand like Maison Noir with strong ChatGPT and Claude presence but a near-zero Gemini score is not unknown to AI systems — it is architecturally underserved on its owned channels. A Black-owned producer with regenerative farming credentials and named vineyard sites who surfaces only in the identity cluster has not failed to build a compelling brand — they have simply not yet published the content that would allow AI systems to place them in four clusters simultaneously.
The window is open now in a way it will not be indefinitely. Ridge's cross-cluster dominance took years to accumulate; the brands that begin publishing strategically in the next 12 months will be the ones AI systems reach for by default as the category consolidates. Identity-forward producers with regenerative farming credentials, named vineyard sites, and accessible occasion-forward positioning are particularly well-positioned — because the content types that drive cross-cluster AI visibility map directly to what these brands already do and know. The gap between what they have and what AI systems can see is almost entirely a publishing and architecture problem.
About This Research
This report is part of an ongoing series examining AI recommendation patterns across premium food, beverage, and hospitality categories. Ally Kiel Consulting publishes original audit data to help founders and operators understand how AI systems currently classify and recommend their brands — and what drives the gaps.
View all research →