Simile, which builds AI simulations of consented real people to predict reactions to

Simile just closed a $200 million Series B at a $2 billion valuation, five months after raising $100 million. Revenue is up five times since the company launched in February. Fortune 100 clients like CVS Health and Deloitte are paying to run AI simulations of real people who’ve consented to have their preferences, behaviours, and likely reactions modelled. The pitch is simple: test a product, a price change, or a campaign against synthetic versions of your actual customers before you commit budget or ship anything.

This isn’t about automating tasks. It’s about de-risking decisions that used to rely on focus groups, surveys, or gut feel. If you’re wondering whether a new SKU will land or a price bump will kill conversion, you can run a simulation instead of rolling the dice in market. According to TechCrunch, enterprises are treating this as strategic insurance, not a nice-to-have.

The speed of the funding rounds and the revenue multiple suggest the model is working. Companies with complex customer bases and high cost of failure are paying premium to get signal before they spend. The consent layer matters here. Simile isn’t scraping or guessing. People opt in, and the simulations are built on that data.

For operators, this is a preview of where AI spend is heading. It’s not all chatbots and content generation. It’s decision support that sits upstream of expensive moves. If you’re running pricing tests, product launches, or market entry plays, this kind of simulation layer is the sort of thing we build into an AI command centre to give you signal before you commit resources. The question isn’t whether synthetic users are real. It’s whether they’re predictive enough to change what you do next.