Simile secures $200 million to examine AI 'twins' of actual consumers.

      A startup is advocating for companies to stop seeking opinions from real individuals and instead rely on artificial ones. Simile has secured $200 million at a $2 billion valuation to create “agentic twins,” AI representations of actual consumers that businesses can survey en masse.

      Greenoaks led the funding round, as reported by the New York Times, with Index Ventures and others participating again. This is a rapid development for the year-old startup, which had raised $100 million less than six months prior.

      Simulating non-existent respondents

      Simile claims that it can forecast human behavior by simulating individuals. Companies utilize this technology to test new products or assess brand perceptions more quickly and affordably than through conventional market research.

      Although the respondents are artificial, the clientele remains genuine. Last fall, CVS Health employed 400,000 agentic twins to determine strategies for encouraging patients to take their medications. Deloitte and the finance application Wealthfront also count among its clients.

      The 💜 of EU tech

      Stay updated on the latest developments in the EU tech landscape, along with insights from our knowledgeable founder Boris and some dubious AI-generated art. It's a free weekly newsletter delivered to your inbox. Sign up today!

      “If you can simulate the world, you can essentially test numerous interventions,” stated Joon Sung Park, Simile’s 32-year-old CEO. He described the technology as “a window into humanity.”

      Real humans created the fakes

      The founders have roots in the research that initiated this area. Park established Simile based on a 2023 Stanford paper on AI agents that replicate human life, akin to an AI version of The Sims. His co-founders include Percy Liang, who contributed to the term “foundation model.”

      The model originates from actual individuals. Simile conducted two-hour interviews with 1,000 representative participants and collaborated with Gallup to incorporate feedback from millions more. It asserts that its predictions are 85% to 99% accurate, depending on the demographic.

      Pricing is based on demand. Simile charges according to the number of agents it simulates, which can range from hundreds of thousands to millions. “The larger the panel, the higher the price,” noted Park.

      A bold and untested concept

      The underlying idea is disputed. The accuracy claims come from Simile itself, and critics question whether AI representations can genuinely reflect human behavior. Competitors like Helm, Artificial Societies, and Aaru are pursuing similar objectives as part of a broader quest for tangible value from AI.

      There is also an element of irony. Simile relies on authentic human data to create its artificial counterparts. This occurs in an environment already saturated with AI agents, where demonstrating one's authenticity has become a business in itself. According to Pulse 2.0, the company aims to model all eight billion people on the planet— a promise that enterprise clients will evaluate based on outcomes.