Proptech AISource: RISMedia
Agents reframe Zestimate & ChatGPT pricing talks — AVMs as opinion, not the number
Original headline: “How Brokers Are Pivoting Pricing Conversations Amid Rise in AI-Driven Valuations” — RISMedia, Oct 23, 2026
RISMedia (Deborah Kearns, Oct. 23, 2026) reports that pricing talks with sellers remain art plus science — comps and AVMs are a useful starting point, not a final valuation, and local knowledge still matters — as AI-driven estimates like Zestimate and chatbot values show up more often at listing appointments. A September study by Finnish and UNC researchers (interviews with 15 real estate professionals) found AVMs can add friction and that agents are changing how they communicate pricing discrepancies to keep trust. Tucson broker Michael Oliver (Oliver Realty) treats AVM, Zestimate, ChatGPT and Claude figures as “another opinion,” while NYC’s Daniel Kramp (Christie’s International Real Estate) gives a range and frames pricing as strategy, and Denver’s Kaylie Curtis (The Gold Standard Brokerage) uses highlighted 30-day comps and a “buyer’s pyramid” visual that listing about 10% below market can capture about 90% of interested buyers versus little interest when overpriced.
Our take (opinion & analysis, not reporting)
For US listing agents: never dismiss a seller’s Zestimate or ChatGPT number — RISMedia’s brokers treat it as another opinion and walk through what AVMs miss (special features, renovations, layout, views, smell/remediation, buyer emotions). Bring recent comps (Curtis prefers the past 30 days, not six-month-old sales), invite sellers to tour comps when useful, and spell out the strategy choice Kramp describes: price high and wait versus price to spark competition. In a slower market Oliver says AVM talks increase as clients filter agents — rehearse a clear local-conditions story before the appointment.
Online discussion: no notable public discussion found yet (checked Oct 24, 2026).