AI Visibility Audit · Fine Dining Series

New York City
Fine Dining Audit

The most AI-contested fine dining market in the country — and what drives it

Published May 2026
Platforms Claude · ChatGPT · Gemini · Perplexity
Prompts Run 50 prompts · 5 clusters · 2× averaged
Location New York, NY

Key Findings

Methodology: Queries were run via API across Claude, ChatGPT, Gemini, and Perplexity — not consumer web interfaces. API responses reflect static training data; consumer-facing products may return different results due to live web access. Each prompt was run twice and results averaged to reduce single-run variance. Brand mentions were extracted using named entity recognition. Results represent baseline AI visibility — the floor, not the ceiling. Entity normalization: Raw NER extraction produced over 1,000 brand variants. Multi-location brands collapsed to primary entity; variant spellings unified; imprint names resolved to canonical restaurant names. Publications, reservation platforms, event venues, and hospitality holding companies excluded from rankings. Market context: 50 prompts returned mentions of over 600 distinct establishments — the largest brand set of any audit in this series.

Platform Divergence —
Top 20 Restaurants

New York City is the most AI-contested fine dining market in the country. The concentration at the top is steep — and the gap between those brands and the rest of the field widens on purchase-intent queries. Across all four platforms, AI visibility is determined less by reputation than by how precisely and repeatedly a restaurant's defining characteristics are described in indexed content.

Restaurant ChatGPT Claude Gemini Perplexity Total
The Modern 397 151 754 59 1,361
Daniel 274 218 222 105 819
Le Bernardin 231 186 144 145 706
Per Se 209 146 97 104 556
Eleven Madison Park 207 149 63 104 523
Masa 137 106 117 43 403
Atomix 81 87 129 78 375
Chef's Table at Brooklyn Fare 68 121 68 55 312
Jean-Georges 153 54 90 6 303
Gramercy Tavern 55 71 47 73 246
Gabriel Kreuther 79 46 56 52 233
Carbone 61 62 52 41 216
Momofuku Ko 133 29 36 13 211
The Grill 55 29 67 52 203
Momofuku 87 32 19 26 164
Manhatta 3 36 50 34 123
One If By Land, Two If By Sea 28 45 23 18 114
Saga 14 10 36 52 112
Cosme 19 15 39 33 106
Lilia 13 33 23 37 106
Highest platform value per row highlighted. Faded values indicate notable platform gaps. The Modern's Gemini total (754) is nearly double its ChatGPT total and more than five times its Perplexity presence — the most extreme platform concentration in the audit. Jean-Georges holds a top-ten overall rank but surfaces only 6 times on Perplexity. Eleven Madison Park underperforms on Gemini (63) relative to every other platform. Momofuku Ko is ChatGPT-dominant — 133 of its 211 mentions come from a single platform. Manhatta is the inverse: nearly invisible on ChatGPT (3 mentions) while performing competitively on the other three.

Strong overall visibility —
near-zero on one platform.

These are not obscure restaurants. They are well-established names whose absence from specific platforms reflects a content indexing gap, not a recognition gap — and the highest-priority content opportunities in this dataset.

Restaurant Other Platform Mentions Gap Platform Gap Mentions
Jean-Georges 297 Perplexity 6
Manhatta 120 ChatGPT 3
Blue Hill at Stone Barns 97 Claude 0
Cote Korean Steakhouse 102 Claude 4
Tatiana by Kwame Onwuachi 82 ChatGPT 2
Atera 87 Claude 4
The Pool 76 ChatGPT 0
Sushi Noz 40 ChatGPT 0
Le Coucou 36 Claude 0
La Grande Boucherie 26 ChatGPT · Claude · Gemini 0
La Grande Boucherie is the most isolated case in this audit: 26 total mentions concentrated entirely on Perplexity, with zero presence on the other three platforms. Blue Hill at Stone Barns and The Pool have complete Claude gaps despite meaningful visibility elsewhere. Tatiana by Kwame Onwuachi has a James Beard Award and extensive national press coverage — and only 2 ChatGPT mentions.

How the category
splits by intent.

NYC fine dining prompts do not return a single consistent restaurant set. AI systems respond differently depending on the diner's intent. Five clusters reveal meaningfully different competitive landscapes — with The Modern leading four of the five.

Cluster 01
Tasting Menu & Chef's Table
Destination meals, serious diners, flagship experience
The Modern Le Bernardin Chef's Table at Brooklyn Fare Masa Atomix

The Modern leads at 309 mentions — a striking result for a restaurant that does not operate a traditional tasting menu format. Its visibility reflects how thoroughly its MoMA setting, Michelin recognition, and Chef Thomas Allen's refined cuisine have been documented. Atomix and Chef's Table at Brooklyn Fare represent the newer generation: both have accumulated deep documentation of their omakase and progressive tasting formats, making them consistently surfaceable against legacy competitors.

Cluster 02 · Highest Single-Cluster Total
Private Dining & Corporate Events
High-spend group bookings, buyouts, corporate entertaining
The Modern Daniel The Grill Le Bernardin Per Se

The Modern's 391 mentions is the highest single-cluster total in the entire audit. Its private dining infrastructure — multiple room configurations, a dedicated events program, Midtown Manhattan location — is well-documented in content AI systems readily index. The Grill's presence reflects its positioning at the historic Four Seasons space in the Seagram Building, generating substantial content around corporate entertaining and power-dining culture.

Cluster 03 · Highest Opportunity
Neighborhood Dining Destination
Local and visitor discovery by neighborhood and occasion
The Modern Le Bernardin Carbone Via Carota Lilia

Carbone, Via Carota, and Lilia represent distinct Manhattan and Brooklyn neighborhoods — Greenwich Village and Williamsburg — with strong community presences that generate locally-specific content AI systems use to answer neighborhood dining queries. This is the cluster where newer and less internationally prominent restaurants have their best opportunity to surface — which makes the gap between the top five and the rest of the field significant for mid-tier brands.

Cluster 04
Special Occasion & Celebration
Anniversary, birthday, proposal, milestone dining
The Modern Le Bernardin Daniel Per Se Eleven Madison Park

The top five are the five most consistently mentioned restaurants in the audit overall — a signal that AI systems have a stable, entrenched map of what "special occasion dining in NYC" means. One If By Land, Two If By Sea underperforms relative to its cultural reputation as New York's most iconic proposal restaurant. Its romantic identity is well-known but not well-documented in the formats AI systems prioritize. Gramercy Tavern's 73 Perplexity mentions confirm its warm, celebration-friendly positioning has been captured in current editorial content.

Cluster 05 · Deepest Chef Identity Signal
Chef & Culinary Identity
Named chefs, international influence, category authority
Daniel The Modern Le Bernardin Jean-Georges Eleven Madison Park

Daniel leads — the only cluster where The Modern does not hold the top position. Daniel Boulud's four-decade presence as a New York culinary figure, his media output, and the volume of chef-attributed content across press and industry sources create a depth of indexed material that is difficult to match. Jean-Georges Vongerichten is similarly positioned, though his Perplexity gap (6 total mentions) is most pronounced in this cluster. Tatiana by Kwame Onwuachi — a James Beard Award winner with extensive press coverage and a clearly documented culinary identity — surfaces almost entirely on Gemini and Claude, with only 2 ChatGPT mentions. Not a recognition gap. A content distribution gap.

Three signal types account for
the majority of high-visibility patterns.

Visibility is not determined by reputation, critical standing, or media budget. It is determined by how precisely and repeatedly a restaurant's defining characteristics are described in content AI systems have indexed.

Signal 01
Chef Attribution

Named chef identity is the single most reliable driver of cross-platform AI visibility. Daniel, Le Bernardin, Jean-Georges, and Eleven Madison Park all benefit from decades of chef-attributed content: interviews, profiles, cookbooks, awards coverage, and culinary philosophy documentation that ties the restaurant's identity to a named individual.

AI systems surface these restaurants not just when asked about fine dining but when asked about specific chefs, culinary movements, and cooking philosophies — expanding their surface area across clusters. Atomix and Tatiana are building this same identity in real time. The gap in their platform coverage reflects how unevenly that documentation has been distributed across the web, not how well-known their chefs are.

Restaurants that have not built chef-attributed content — where the brand identity is the space or the occasion rather than the kitchen — have a narrower surface area in AI queries. They surface in occasion-based clusters but rarely in chef-driven or culinary identity searches.
Signal 02
Multi-Cluster Documentation

The restaurants with the highest total visibility are documented across multiple use cases, not just one. The Modern's outsized visibility is partly explained by the breadth of its documented use cases: tasting menus, private events, MoMA-adjacent tourism, corporate entertaining, special occasions. Each use case generates distinct content that makes the restaurant surfaceable across all five query clusters.

Le Bernardin and Daniel are similarly multi-cluster because their content spans chef identity, private dining, occasion dining, and culinary authority. Restaurants that are primarily documented in one register — romantic proposal venues, or neighborhood spots — surface in the relevant cluster but disappear from the others, limiting their total visibility ceiling.

Single-use-case documentation creates single-cluster visibility. A restaurant known only for proposals, or only for its tasting menu, can only be surfaced by queries that match that one identity.
Signal 03
Current Editorial Presence

In New York, recency of coverage is a material factor in Perplexity visibility — and Perplexity is a bellwether for the direction the other platforms are moving. Saga, Cosme, and Lilia all perform above their overall rank on Perplexity — a pattern consistent with restaurants generating ongoing editorial coverage, not just legacy documentation.

Jean-Georges and Masa represent the inverse: strong historical coverage but limited current web presence in the formats Perplexity prioritizes. As AI platforms increasingly incorporate live web access, the distinction between legacy reputation and current content presence will widen. Restaurants generating fresh editorial content now are building the asset that will matter most in the next generation of AI-driven discovery.

Legacy documentation gets a restaurant into the training data floor. Current editorial presence determines its ceiling as platforms move toward real-time retrieval.

The Modern's 1,361 total mentions are real — but 754 of them come from Gemini alone. On Perplexity, it would rank outside the top ten. The highest total in this audit is also one of the most platform-dependent. A 13:1 Gemini-to-Perplexity ratio is a concentration risk, not a strength.

These are not reputation gaps.
They are content gaps.

Daniel, Jean-Georges, Blue Hill at Stone Barns, and Tatiana by Kwame Onwuachi are not poorly regarded restaurants — they are some of the most critically recognized dining establishments in the country. What this audit measures is the gap between a restaurant's reputation and its content footprint: the degree to which the characteristics that make it relevant to a specific query have been precisely documented in formats AI systems can retrieve and surface.

Three content types moved the needle most consistently. Chef-attributed culinary philosophy drives visibility in the Chef & Culinary Identity cluster and compounds across all others. Private dining and event documentation — room configurations, capacity, menus, inquiry processes — directly addresses the highest purchase-intent cluster. And occasion-specific positioning content explains why One If By Land, Two If By Sea surfaces far below its cultural reputation in special occasion queries: the romantic identity is well-known but not well-documented in the formats AI systems prioritize.

Research published at KDD 2024 found that applying generative engine optimization techniques improved AI citation rates by an average of 40% — and that lower-ranked sources saw the largest gains, with some improving by as much as 115%. The field is not locked. AI visibility is not purely a function of legacy reputation or media budget.

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 →