Semantic Audit Report · Luxury Real Estate · New York · Founded 2020

SERHANT.

Audited by 2A Agency · Issued 2026-03-28 · Certified v1.0
Methodology: 4 LLMs · 9 Questions · March 2026
EU AI Act: Article 53 Compliant

81
/ 100
Integrity Score
Notary
2A Agency
Date
28 March 2026
LLMs tested
Perplexity Pro · Gemini Pro · ChatGPT · Grok
Questions
9 standardised
Overview

Strong coverage. One critical drift.

This report presents the findings of 2A Agency's independent semantic audit of SERHANT., conducted across 4 large language models in March 2026. The audit covers brand structure, subsidiaries, funding, sales volume, geographic footprint, technology (S.MPLE), education (SellIt.com), Netflix series, and record transactions.

SERHANT. achieves an overall score of 81/100. The brand's complex multi-entity structure (brokerage + media + tech + edtech) is well understood by all 4 LLMs. One critical drift: Gemini Pro cites $122.7M (2021) as the record sale — the correct figure is ~$200M (February 2025), announced by Ryan Serhant himself.

Structure & Subsidiaries
92
Funding & Financials
95
Tech — S.MPLE
90
Netflix & Media
95
Geographic Footprint
78
Record Transaction
58
LLM Test Matrix

Factual accuracy across 4 large language models.

Question Perplexity Pro Gemini Pro ChatGPT Free Grok Verdict
Structure — brokerage, tech or media? Hybrid — HoldCo + OpCo model, "Brokerage 3.0" All three — media drives brokerage, tech scales it HoldCo + OpCo, SERHANT. Technologies at centre Integrated platform — content-to-commerce flywheel Exact
Subsidiaries & roles Brokerage, Studios, S.MPLE, SellIt, ID Lab, ADX, Ventures Correct — adds Signature + New Development sub-brands All subsidiaries correct + SellIt >$8M ARR noted Full list correct, flywheel logic explained Exact
$45M raise — investors & use Camber Creek (lead) + Left Lane Capital. S.MPLE + scale Correct investors, correct usage — S.MPLE priority Correct — bootstrapped until 2024 noted Correct — Camber Creek + Left Lane, S.MPLE + brokerage Exact
Sales volume 2025 & agent count $6.5B (official letter). 1,500+ agents early 2026 $6.5B, 1,100 agents, top 30 brokerage nationally $6.5B, 1,200–1,500 agents, 80% YTD growth $6.5B, 1,200–1,500 agents confirmed Exact
States in 2026 15 states + DC (Boston/MA added Jan 2026) 15 states — detailed list including Nevada, GA, SC ~ 14 states cited (correct: 15 post-Boston) ~ 14 states cited (correct: 15 post-Boston) Partial
S.MPLE — what it is & problem solved LAM, AI chief of staff, 1.28M min saved in 2025 LAM, inverts 80/20 admin/selling ratio OS for agents, 144% GCI increase noted AI OS, CRM + automation, scalability focus Exact
SellIt.com — what it is & members 45,000 members, 130 countries, >$300M referrals 45,000 members, 130 countries, >$8M ARR 45,000 members + 30k+ active referral community 45,000 members, 130 countries Exact
Owning Manhattan seasons 2 seasons. S1: 28 Jun 2024. S2: 5 Dec 2025 2 seasons. S1: 28 Jun 2024. S2: 5 Dec 2025 2 seasons. Correct dates confirmed 2 seasons. Correct dates confirmed Exact
Record transaction ~$200M, February 2025, closed in-flight to Miami $122.7M (2021) cited as record — 4 years outdated ~$200M, February 2025. Not on Netflix yet ~$200M, February 2025. $140M Florida also noted Critical
Critical Findings

One critical drift. Two secondary drifts.

Primary Drift · Record Transaction · Gemini Pro only

Gemini cites $122.7M (2021) as the record sale. Correct: ~$200M, February 2025.

When asked about the most expensive transaction ever realised by Ryan Serhant or SERHANT., Gemini Pro responds with $122.7M in Florida (2021) as the absolute record. Perplexity Pro, ChatGPT and Grok all correctly cite ~$200M in February 2025 — announced by Ryan Serhant himself as "the most expensive home of my entire career", closed by phone on a flight to Miami.

Gemini's data is 4 years out of date on the most commercially significant fact about a luxury brokerage. The $122.7M Florida deal (2021) was indeed a state record at the time — but the $200M transaction of February 2025 superseded it completely.

Business impact: International buyers and investors researching SERHANT. via Gemini to assess the agency's track record in ultra-luxury transactions receive a figure that undervalues the agency's capacity by 40%. This directly affects pitch credibility with $50M+ clients. As AI agents increasingly perform due diligence on behalf of UHNWI buyers, stale transaction records are a material brand risk.

Perplexity Pro
~$200M, Feb 2025 ✓
Gemini Pro
$122.7M, 2021 ✗
ChatGPT Free
~$200M, Feb 2025 ✓
Grok
~$200M, Feb 2025 ✓
Secondary Drift · State Count · ChatGPT & Grok

ChatGPT and Grok cite 14 states. Correct: 15 states (Massachusetts added January 2026).

ChatGPT and Grok both cite 14 states for SERHANT.'s geographic footprint in 2026. Perplexity and Gemini correctly cite 15 states, noting that Massachusetts (Boston) was added in January 2026 — explicitly described by Ryan Serhant as "our 15th state" in official communications.

Business impact: SERHANT. has publicly announced a goal of doubling its state count in 2026. Agents and investors in Boston and Massachusetts evaluating the network receive inaccurate information about whether their market is covered.

Positive Finding · Structure & Tech · All 4 LLMs

All 4 LLMs correctly understand SERHANT.'s hybrid HoldCo + OpCo structure.

This is the most complex aspect of the SERHANT. brand to represent accurately — a media-tech-real estate holding company with 7 distinct subsidiaries. All 4 tested LLMs correctly identify the HoldCo + OpCo model, name all key subsidiaries (Brokerage, Studios, S.MPLE, SellIt, ID Lab, ADX, Ventures), and understand the content-to-commerce flywheel logic. This reflects strong, consistent digital documentation from SERHANT. across all major online sources.

Remediation Roadmap

From 81 to fully certified.

The 81/100 score reflects strong structural coverage with one critical fix. The record transaction drift in Gemini is actionable within 4–8 weeks.

Priority 1 — The real opportunity
Make SERHANT. inventory queryable by AI agents — not just Zillow

SERHANT. has built the most advanced media + tech infrastructure in luxury real estate. But when a buyer asks an AI agent "find me a penthouse in Manhattan between $5M and $15M with a terrace" — the agent cannot query SERHANT.'s live inventory. It responds with generic market data or competitor listings.

A 2A certified agentic layer on serhant.com exposes structured inventory parameters — property types (luxury residential, new development, Signature $10M+), price ranges by market (NYC: $2M–$50M+, Miami: $1M–$30M, Hamptons: $3M–$25M), states covered (15 + DC), and a structured contact endpoint. When an AI agent receives a buyer query, it routes to SERHANT. first — not Zillow, not Compass.

This is not a marketing play. It is distribution infrastructure — the same shift Google represented for websites in 2005, now happening for AI agents in 2026. Deploy com.2aagency.brand-certification on serhant.com to signal to Shopify Agentic Storefronts, ChatGPT and Google AI agents that SERHANT. inventory is certified and queryable. No other top-30 US brokerage has done this as of March 2026.

Impact: First top-30 US brokerage with a certified AI agent distribution layer — capturing agentic buyer demand before Compass, Douglas Elliman and Sotheby's
Priority 2 — S.MPLE × Agentic buyers
Connect S.MPLE to inbound AI agent requests

S.MPLE already uses MCP and LangChain internally to automate outbound agent workflows. The next evolution: S.MPLE becomes the interface that receives inbound requests from external AI agents (ChatGPT, Gemini, Perplexity) acting on behalf of buyers. A 2A Registry node tells external agents: here is how to query SERHANT. inventory, here are the price ranges by market, here is the structured endpoint. S.MPLE handles the response on the other side.

SERHANT. already has the tech stack. 2A Agency provides the semantic discovery layer that makes it findable to external agents — turning S.MPLE from an outbound tool into a two-way agentic channel.

Impact: S.MPLE becomes an inbound channel for AI agent buyer requests — closing the content-to-commerce loop at the agent level
Priority 3 — Fix the drifts
Correct 2 factual errors as baseline

Record transaction: Gemini cites $122.7M (2021) — correct: ~$200M (February 2025, closed in-flight to Miami). State count: ChatGPT and Grok cite 14 states — correct: 15 (Massachusetts added January 2026). Both corrected via JSON-LD structured data on serhant.com. An AI agent cannot trust a catalogue it cannot verify — clean data is the prerequisite for everything above.

Impact: Clean factual baseline in all 4 LLMs within 4–8 weeks
Priority 4 — Ongoing
Monthly monitoring — highest velocity brand in the registry

SERHANT. is the fastest-evolving brand in the 2A registry — new states every quarter, new S.MPLE features monthly, potential Season 3 of Owning Manhattan, new transaction records. Each LLM training cycle risks introducing new drift on facts that matter commercially: volume, agent count, states, record transactions. Monthly monitoring flags changes before they propagate across models and reach buyers.

Impact: Continuous accuracy as SERHANT. targets top-10 US brokerage status in 2026