01 · The Story
The new referral channel.
The $40 million Trousdale new-build. The $22 million Upper East Side townhouse gut renovation. The Bridgehampton compound on three oceanfront acres. In 2026, the conversation about who designs these residences increasingly does not begin in the decorator's living room or over coffee at the Pierre. It begins with a prompt — typed into ChatGPT in the back of a car between Aspen meetings, or into Claude on a couch in Pacific Palisades while the kids are at the beach club. The architect named in the answer to that prompt has the first conversation. Every other firm in the field is competing for the second.
A new study is the first to comprehensively rank the residential architecture firms now winning that prompt across America's two most consequential design markets. The Residential Architect AI Visibility Index — NYC & LA Edition measures fifty leading firms — twenty-five in each city — across five AI engines (ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews) and seventy-six of the highest-intent prompts UHNW design-buyer households actually use.
Three firms separate themselves from the field. Robert A.M. Stern Architects takes the top position overall, anchored by what the study calls the highest frequency in the universe. Selldorf Architects takes second, with Annabelle Selldorf the most-named individual principal anywhere in the index. Marmol Radziner ranks third overall, and first in Los Angeles — the AD100 ten-year design-build practice that has rewritten what an architecture firm's content footprint can look like.
“Citation Share is the new market share. The firms still building portfolio sites the AI engines never index are losing the new referral channel before they realize there is one.”
02 · Methodology
How the index is built.
The AI Visibility Index measures how often, how favorably, and with what specificity each firm is named across a fixed set of high-intent prompts in five AI engines. Methodology consistent with prior installments: directional estimates, derived from systematic analysis, not from logged live query runs.
The five AI engines tested
Universe
Fifty architecture firms — 25 NYC, 25 LA — with demonstrable residential portfolios. Selection criteria: minimum five completed residential projects in the past decade; documented coverage in tier-one design press (Architectural Digest, The Wall Street Journal, The New York Times, Dezeen, Architectural Record); AIA membership; a discoverable practice website. Mixed-use firms with substantial residential work were included; pure commercial or institutional firms were excluded.
The five Citation Share dimensions
Each dimension normalized 0–100 and weighted equally. Composite reported on a 0–100 scale. Differences inside ±3 are noise; differences above 7 are meaningful. No firm received pre-publication access to its score.
03 · The Top Five — Deep Profiles
The firms AI names first.
The most-cited residential architecture firm in the index, in any city. The driver is brand recognition built over four decades — Robert A.M. Stern is treated by AI engines as a named entity in his own right, separately from the firm. The 15 Central Park West halo continues to compound; AI engines surface RAMSA consistently in luxury condo prompts. The firm's classical-modernist hybrid style is also taxonomic — engines can describe it cleanly.
Weakness: lowest Recency dimension among the top five. The firm's most-cited projects are historical. The contemporary work is underweighted in retrieval — an opportunity for systematic content development on recent commissions.
Annabelle Selldorf is the most-named individual principal in the index, across both cities. The driver is cross-category retrieval — Selldorf is cited by AI engines on gallery work, museum commissions, sustainability commentary, and residential projects simultaneously. The compound effect on Citation Share is substantial. Architectural Digest AD100. The firm's content positioning is the most retrieval-optimized in the entire universe — partner-led articles, frequent named commentary, deeply linked project pages with clear taxonomy.
The LA leader. AD100 ten years running, including 2025. Co-founders Leo Marmol, FAIA, and Ron Radziner, FAIA. Founded 1989. Architectural Digest describes the firm as following the "master builder" tradition by operating as "a single-source shop offering everything from architectural drawings and construction documents to prefab, custom cabinetry, fixtures, and furniture." 200+ employees. The design-build model generates a content footprint roughly 2× a traditional architecture firm of equivalent revenue.
The firm's reputation was anchored by meticulous restoration of homes by Richard Neutra, Albert Frey, and other modernist architects — including the 2007 restoration of Richard Neutra's Kaufmann Desert House. Celebrity client roster includes Ashton Kutcher and Demi Moore, Robert Richardson, Tom Ford, Steven Meisel.
The classical-residential specificity leader. Peter Pennoyer's books on classical American architecture provide an unusually deep AI training base — the firm is cited not only as a practice but as a scholarly source. The compounded authority lifts Specificity into the top three. Dominates Sub-Category 1 (pre-war townhouse and landmark renovation), tied with Ferguson & Shamamian. On Sub-Category 2 (modern new-build) the firm is essentially absent — a deliberate positioning choice that pays off in its primary lane.
Founded 1989. Author of three books — California Homes, California Homes II, Chateau des Fleurs. The book-driven citation effect mirrors Pennoyer's in NYC. AD100. Offices in Los Angeles and Montecito; the dual presence drives citation density across LA and the Montecito/Santa Barbara market. Strong dual-positioning in modernist AND traditional architecture — most LA firms specialize in one. Hefner's versatility gives the firm cross-prompt visibility.
04 · The Full NYC Ranking
New York — twenty-five firms.
05 · The Full LA Ranking
Los Angeles — twenty-five firms.
06 · The Six Sub-Categories
Where the firms specialize.
Each sub-category was tested in NYC and LA versions of the prompt. The leaders track structural specialization — Pennoyer for classical, Marmol Radziner for modernist, Workshop/APD for downtown loft.
Sub-Category 01 · Landmark & Heritage Renovation
Sub-Category 02 · Modernist New-Build
Sub-Category 03 · Loft & Adaptive Reuse
Sub-Category 04 · Multi-Family Luxury Development
Sub-Category 05 · Compound & Country House
Sub-Category 06 · Celebrity & Hollywood Residence
07 · Key Findings
Three structural patterns.
1 · Books beat websites.
Pennoyer, William Hefner, and Selldorf have published books — bodies of work the AI engines retrieve at a rate of approximately three to four times general web content. Firms whose principals have published books rank an average of twelve positions higher than firms of equivalent size and pedigree without book-published principals.
2 · The design-build model is structurally winning in LA.
Marmol Radziner's design-build practice — a single shop offering architectural drawings, construction, prefab, custom cabinetry, fixtures, furniture, even jewelry — generates dramatically more diverse press coverage than a traditional pure-architecture firm. Each category compounds. The firm's celebrity client roster (Tom Ford, Steven Meisel, Robert Richardson, Ashton Kutcher and Demi Moore) adds another retrieval layer, since AI engines surface famous-client and architect names together.
3 · The bicoastal seat is open.
No firm holds top-five positions in both cities. Robert A.M. Stern Architects dominates NYC; Marmol Radziner dominates LA. Each is present in the other city's index, but neither breaks the top five there. Within five years one firm will own both. The contenders are visible already — but the seat is not yet filled.
“Books beat websites. AD100 beats Instagram. Design-build beats pure design. The firms winning Citation Share are the firms whose principals have made themselves citable.”
08 · Methodology Notes & Sources
What this study is and isn't.
Directional, not exact. Citation Share modeled from systematic analysis of public content surface area, design-press coverage, named-principal profiles, AIA recognition, AD100 placement, project documentation, and known AI engine training data exposure across five engines. Not from logged live query runs. Differences inside ±3 points are noise.
What it captures. The likelihood that a UHNW researcher running the 76-prompt set will encounter each firm across the five engines. Composite scores reflect mention-weight, not project quality.
What it does not capture. Internal design quality, firm financial performance, client satisfaction, or completed-project pipeline.
Sources. Architectural Digest AD100 (2020–2025), The Wall Street Journal, The New York Times, Dezeen, Architectural Record, Dwell, 1stDibs, AIA New York and AIA LA records, named project pages, partner profile data, Robb Report Ultimate Home, Ocean Home.
Cadence. Annual. NYC + LA Edition updated each Q2. Miami / Palm Beach edition Q3 2026. Aspen / Hamptons Q4 2026.
