Simile Raises $200M to Simulate Human Behaviour With AI

What if a company could test a product, policy or message on a simulated population before trying it on real people?

What happened

Simile has raised a $200 million Series B led by Greenoaks at a $2 billion valuation. The round follows a $100 million Series A announced only five months earlier, giving the company an unusually fast funding trajectory.

Its platform builds AI-based simulations of groups of people and estimates how those groups might react to products, policies or business decisions. The pitch is not simply another chatbot. Simile wants to model patterns of behaviour at population level, creating a faster preliminary layer before companies run surveys, focus groups or live experiments.

The company is still young, and the central claim—that simulated people can reliably predict the behaviour of real ones—has not yet been independently proven across a broad range of decisions. That limitation is especially important when the output could influence healthcare, public policy or high-stakes commercial choices.

Why it matters

Traditional research can be slow and expensive. If simulations are directionally accurate, companies could test more ideas, reject weak options earlier and reserve real-world research for the most promising ones.

But a convincing interface can also hide weak assumptions. A model trained on incomplete or biased data may produce precise-looking answers that fail when exposed to real people.

The bigger picture

AI is moving from generating content to modelling possible futures. Digital twins already simulate machines, factories and supply chains; Simile is extending that idea to human behaviour. The commercial opportunity is huge, but so is the validation burden. In this category, predictive accuracy—not fluency—will determine whether the product becomes infrastructure or an expensive storytelling tool.