- Startup Simile is developing “agentic twins” technology—AI agents modeled after real people to answer surveys and predict customer behavior.
- The company raised $100 million in a Series A round led by Index Ventures, showing significant investor interest in behavior-simulating AI.
- Simile spun off from Stanford University in 2024 after the founding team published research on simulating human behavior using AI.
- Simile’s technology creates “digital clones” of real people by interviewing them about their personalities and preferences, combined with shopping behavior data.
- These AI agents can then be asked thousands of questions to predict consumer reactions in various situations.
- Customers can access a library of these agents to conduct market research faster and cheaper than traditional surveys.
- Service costs range from approximately $150,000 to several million dollars per year for each business.
- Healthcare giant CVS Health has used “agentic twins” to answer questions about customer behavior instead of real-person survey groups.
- The CVS system was built from 2.9 million survey responses from over 400,000 consented participants.
- In tests, the AI twins could replicate known market research results with up to 95% accuracy.
- The company plans to expand the system to over 100,000 AI agents to test store layouts, product designs, and marketing messages.
- Opinion research organization Gallup is also partnering with Simile to provide over 1,000 “digital twins” for customers to conduct policy, social trend, and corporate research.
- AI agents can also be used in other fields such as clinical trials, focus groups, and consumer behavior simulations.
- However, experts like Gartner analysts warn that this technology is still in its early stages and cannot completely replace data collected from real humans.
- Therefore, businesses still need to verify results with actual surveys to avoid bias or “AI hallucinations.”
📌 Simile’s “agentic twins” technology is opening up a new way to conduct market research by simulating human behavior with AI. With data from 400,000 people and 2.9 million survey responses, CVS’s system can replicate research results with up to 95% accuracy. Thanks to the ability to ask unlimited questions and operate continuously, AI can reduce costs and accelerate research. However, experts believe that data from real humans is still necessary to verify and ensure reliability.
