Research · May 2026
NYC AI Visibility Benchmark: May 2026
Peloton scores 0.0. Skadden Arps scores 33.3. We scanned 30 New York brands across B2B SaaS, legal services, and fitness — and found that category, not brand size, determines who AI recommends.
Peloton spent over $100 million on brand marketing between 2020 and 2023. The company has one of the most recognizable brand identities in consumer fitness. It has a global content operation, a podcast, a library of thousands of instructor-led classes, and a loyal customer base that generates organic word-of-mouth at scale.
When we asked Perplexity "what's the best home workout equipment for someone who wants structured classes," Peloton did not appear.
When we asked "best fitness options for busy New Yorkers who want premium workout experiences," Peloton did not appear.
Across 12 buyer-intent fitness prompts, Peloton appeared zero times. AI Brand Score: 0.0.
This is not a data anomaly. It reflects something structural about how AI engines recommend brands — and it is the central finding of this study.
Disclosure: Envoyra sells the AI visibility scanning tool used to collect this data. We have a financial interest in findings that demonstrate the value of AI visibility measurement. The methodology is disclosed in full below. Underlying scan data is available to clients and press on request.
The study
We scanned 30 New York City brands across three verticals in May 2026: 10 B2B SaaS companies, 10 legal services firms, and 10 fitness brands. Each brand received 12 buyer-intent prompts on Perplexity sonar-pro. Prompts were generated using brand-specific intake: industry, buyer persona, target use case, location context. Total prompt-response pairs: 360.
This study is part of a larger Envoyra benchmark dataset now covering 45 brands across 4 verticals and 2 cities (DC and New York). Results are reported as AI Brand Score: a 0–100 composite weighted across engines, with Perplexity as the primary signal source for this study.
360 | Prompt-response pairs analyzed — 30 NYC brands × 12 buyer-intent prompts, May 2026
The scores
B2B SaaS — avg AI Brand Score: 48.3
| Brand | Score | Top Competitor in AI |
|---|---|---|
| Datadog | 83.3 | Amazon CloudWatch |
| Braze | 75.0 | Intempt |
| Pager | 66.7 | Datadog |
| MongoDB | 66.7 | Google BigQuery |
| Riskified | 66.7 | Stripe Radar |
| Attentive | 50.0 | Klaviyo |
| Yotpo | 25.0 | HubSpot |
| Movable Ink | 16.7 | HubSpot |
| Sprinklr | 16.7 | HubSpot |
| Conductor | 16.7 | HubSpot |
Legal Services — avg AI Brand Score: 25.8
| Brand | Score | Top Competitor in AI |
|---|---|---|
| Cleary Gottlieb | 33.3 | Sullivan & Cromwell |
| Sullivan & Cromwell | 33.3 | Latham & Watkins |
| Davis Polk | 33.3 | Latham & Watkins |
| Skadden Arps | 33.3 | Sullivan & Cromwell |
| Weil Gotshal | 25.0 | Holland & Knight |
| Greenberg Traurig | 25.0 | Latham & Watkins |
| Willkie Farr | 25.0 | Holland & Knight |
| Kramer Levin | 16.7 | Holland & Knight |
| Epstein Becker Green | 16.7 | Latham & Watkins |
| Proskauer Rose | 16.7 | Wachtell Lipton |
Fitness — avg AI Brand Score: 20.8
| Brand | Score | Top Competitor in AI |
|---|---|---|
| SoulCycle | 66.7 | Equinox |
| Equinox | 58.3 | Mercedes Club |
| Rumble Boxing | 50.0 | Central Park Boxing |
| Barry's | 25.0 | Equinox |
| Shadowbox | 8.3 | Barry's |
| Tone House | 0.0 | — |
| AARMY | 0.0 | — |
| Peloton | 0.0 | — |
| Fhitting Room | 0.0 | — |
| ClassPass | 0.0 | — |
Finding 1: Peloton is invisible — and that's the point
The Peloton finding is not a story about Peloton specifically. It is a story about what AI visibility measures and what it does not.
Peloton has extraordinary brand awareness. If you asked a thousand people in New York to name a fitness brand, most would say Peloton. But brand awareness is aided recall — it measures whether people recognize a name they already know. AI recommendations are unaided — they fire when a buyer doesn't yet know which brand to choose and is asking for guidance.
Peloton's 0.0 score means it is not appearing in the exact queries where undecided buyers are being made. Buyers who already know Peloton is right for them will find it. Buyers who don't know yet — the acquisition frontier — are being directed elsewhere by AI.
Brand awareness measures aided recall. AI visibility measures unaided discovery. Peloton has one and not the other. Most brands don't know which side of that line they're on.
ClassPass scores 0.0 for the same structural reason. ClassPass has a strong consumer brand, significant PR coverage, and a recognizable product. But it is a platform, not a fitness category. When buyers ask "best boutique fitness classes in NYC," AI recommends specific studios — SoulCycle, Equinox, Barry's — not the aggregator. ClassPass's entire positioning works against it in AI recommendations.
Finding 2: B2B SaaS splits at the infrastructure layer
The B2B SaaS vertical shows the clearest pattern of any vertical in this study.
Infrastructure and developer tools dominate: Datadog (83.3), MongoDB (66.7), Riskified (66.7). Marketing and engagement platforms cluster at the bottom: Movable Ink (16.7), Sprinklr (16.7), Conductor (16.7).
The gap is not about company size or marketing budget. Sprinklr is a publicly traded company. Movable Ink raised over $100M. Both score below 17.
66.7pp | Gap between Datadog (83.3) and Movable Ink (16.7) — same city, same budget tier, opposite AI visibility outcomes
The explanation is category specificity. Datadog is "cloud monitoring and observability." When a buyer asks "how do I monitor my cloud infrastructure," the answer is structurally clear and Datadog owns it. Sprinklr is "unified customer experience management" — a category description that AI cannot confidently match to a specific buyer query. The vaguer the category, the harder AI works to name you.
HubSpot appears as the top AI-recommended competitor for Yotpo, Movable Ink, Sprinklr, and Conductor — four different NYC companies in four different subcategories of marketing technology. HubSpot's AI dominance is not because it outcompetes each of these tools in their specific niche. It is because HubSpot has invested years in content that answers the exact questions buyers ask before they know which tool they need.
Finding 3: Legal is a category waiting for a first mover
Every law firm in this study scores between 16.7 and 33.3. The range is 16.6 points across 10 firms — the most compressed distribution of any vertical we have studied.
Skadden Arps (33.3) and Cleary Gottlieb (33.3) are among the most prestigious law firms in the world. They appear in AI recommendations at the same rate as Davis Polk and Sullivan & Cromwell. The firms at the bottom — Kramer Levin (16.7), Proskauer Rose (16.7), Epstein Becker Green (16.7) — are significant practices with decades of case history and client press.
No firm has broken out. No firm has figured out AI.
Latham & Watkins — a Los Angeles-headquartered firm — appears as the top AI-recommended competitor for three NYC-primary firms. Being headquartered in the city your clients are in does not matter to AI. Having content that answers their questions does.
The opportunity here is significant and first-mover-exclusive. The first NYC law firm to build structured AI-visible content — practice area pages that answer specific questions buyers ask before they hire a lawyer, not just credential listings — will pull away from a compressed field. The category average is 25.8. A firm that moves to 50+ will have a visible, measurable AI advantage over every competitor in this study.
Cross-vertical context: where NYC sits vs. DC
Combined with our April 2026 DC study (15 DTC wellness and beauty brands), we now have a 45-brand dataset across 4 verticals and 2 markets.
| Vertical | City | Avg Score | Highest | Lowest |
|---|---|---|---|---|
| B2B SaaS | NYC | 48.3 | Datadog 83.3 | Movable Ink 16.7 |
| Legal Services | NYC | 25.8 | Cleary Gottlieb 33.3 | Kramer Levin 16.7 |
| Fitness | NYC | 20.8 | SoulCycle 66.7 | Peloton 0.0 |
| DTC Wellness | DC | ~22.0 | Take Care 41.7 | Multiple 8.3 |
Three of four verticals cluster between 20 and 26. B2B SaaS is the outlier — and within it, the split between infrastructure tools and marketing platforms is what drives the average up.
The consistent finding across all four verticals: every category has a small number of brands with strong AI presence (SoulCycle, Datadog, Zevia, Braze) and a large number that are functionally invisible to buyers in discovery mode. The leaders are not necessarily bigger or better-funded. They have content that answers the specific questions buyers ask before they know which brand to choose.
What separates the visible from the invisible
Across 45 brands and 4 verticals, the structural pattern is consistent.
Visible brands have content that answers pre-purchase questions at the category level — not brand-level. Datadog doesn't just describe what Datadog does. It answers "how do I monitor cloud infrastructure?" SoulCycle doesn't just describe SoulCycle. It is present in the content that answers "what are the best fitness classes in New York?"
Invisible brands have content that describes themselves to people who already know them. This is brand marketing. It is not AI visibility.
The three highest-leverage actions based on this dataset:
1. Map your prompts before you create content. The prompts that matter are not "what is [your brand]?" They are the questions buyers ask when they are choosing a category solution and don't yet have a brand in mind. Run a scan to see which prompts you appear on and which you don't. The gaps are your content brief.
2. Target third-party citation sources, not your own site. Latham & Watkins dominates legal AI recommendations without being in this city. HubSpot dominates marketing software recommendations for categories it doesn't specifically own. Both are heavily cited in third-party publications, directories, and review platforms that AI engines trust. Your own website is not the primary citation source. The publications that cover your category are.
3. Specificity beats volume. The legal firms in this study have more published content than almost any other business category. Press releases, case announcements, attorney bios, practice area descriptions, client alerts. None of it moves the AI needle because none of it answers the pre-purchase question a buyer is actually asking. One page that answers "what kind of law firm do I need for a cross-border M&A deal?" will outperform 100 attorney bio pages for AI discovery purposes.
Methodology
Scan date: May 2026. Brands: 30 NYC-area brands across B2B SaaS (10), Legal Services (10), and Fitness (10). Engine: Perplexity sonar-pro (live web retrieval, 100% citation density). Prompts: 12 buyer-intent prompts per brand, generated via Claude using brand-specific intake including industry, buyer persona, location (New York City), and use case. Scoring: AI Brand Score — Share of Model on Perplexity, expressed as percentage of 12 prompts where brand appeared. Total prompt-response pairs: 360. Combined dataset: 45 brands across DC and NYC, 4 verticals, 540 prompt-response pairs.
Single-shot Perplexity scores are directional signals. Emerging-tier brands (under 20%) should be interpreted as floor estimates with ±3–5pp variance on re-scan. Trend measurement over 4+ weeks is recommended before drawing strategic conclusions.
Prior study: DC-Area Brand AI Visibility Benchmark, April 2026 — 15 DTC wellness and beauty brands, 4 engines, 336 prompt-response pairs.
See where your brand stands
Most brands in this study were measuring their AI visibility for the first time. The findings are not unique to New York or to these specific companies. They reflect how AI recommendation engines work — and most brands have not built for it yet.
The research documents which buyer-intent prompts brands appear on, which they miss, and which competitors are being recommended instead.