Simile's $2 Billion Bet: Synthetic Audiences Reshape Marketing

Simile raised $200 million at a $2 billion valuation in July, roughly five months after its Series A, to build AI simulations of real consumers that companies can survey. The round signals how much investors are betting on synthetic audiences — simulated consumer groups built from customer records, interviews and demographic profiles — as marketing's next research tool.

Adobe is already in the game. Its Brand Intelligence product lets marketers test creative and messaging against AI-generated customer profiles inside the company's own stack.

The economics are straightforward. Marketing departments face pressure to produce more, spend less and improve quality simultaneously. Synthetic audiences let a team present work to simulated consumers, examine predicted responses, revise and try again — without recruiting another human panel. For companies that understand their customers and have validated predictions against actual behavior, that's enormously useful.

What does 'we tested it' actually prove?

The appeal goes beyond efficiency, the analysis argues. Synthetic audiences also deliver something invaluable: a defensible reason to proceed. "We tested it, and this is what customers preferred" carries different authority from "we think this is a good idea" — even when the customers are simulated.

Steven Waterhouse, in his essay Learning Machines All the Way Down, frames the broader shift: AI reduces the cost of trying things, examining results and adjusting. Companies organized around that loop can fold what they learn into the next decision, improving and accelerating as they go.

United States Air Force Colonel John Boyd formalized this process as the OODA loop: Observe, Orient, Decide, Act. Businesses often cite it as an argument for speed. Boyd's version gives considerable weight to orientation — how we understand what's happening, drawing on experience, culture and new information. In his diagram, a decision is a hypothesis and an action is a test.

Here lies the catch. Synthetic audiences can work through understanding before committing resources. They may expose an unnoticed assumption or suggest an unconsidered interpretation. But another simulated response doesn't add another observation of how people actually behave. A team can run hundreds of iterations, growing increasingly confident in a campaign, without learning whether the audience shares that confidence.

What happens when prediction drives approval?

Boyd also wrote that "adaptability implies variety and rapidity," tying both to the ability to be unpredictable and cope with unforeseen circumstances. Applied to creative work, that asks more than producing variations quickly — it requires preserving a range of possibilities broad enough to discover something otherwise unattempted.

David Perell recently described the counter-dynamic on TBPN, crediting Derek Thompson with the observation that Hollywood has succumbed to the "Moneyball-ification of our world." In Perell's telling, the industry has become boring partly because it has become intelligent about risk. "The problem is you end up getting Spider-Man 74 or Lord of the Rings 7 rather than a new show."

A studio can make a succession of reasonable decisions and still produce a culture in which unfamiliar stories struggle to get made. The same logic threatens marketing teams.

A creative team doesn't need an instruction to obey the model. If a favorable score helps secure approval, members will notice which proposals move forward easily and which require lengthy defense. Each accommodation may be minor; over time they narrow the range of work the team produces. Success within the process can gradually become the team's working definition of good creative judgment.

Waterhouse describes a related development in the "Mirror Image" section of his essay: we build intelligence in our likeness, then reorganize our institutions around its image. In a creative department, that reorganization happens inside individual habits. We adapt to the tool until its preferences feel like our own judgment.

Isaac Asimov dramatized the risk in Foundation and Empire. Hari Seldon's psychohistory forecasts the behavior of vast populations — until the Mule, a mutant who alters emotions and loyalties, throws the Plan off course. An individual outside the model's assumptions changes the future it was supposed to predict.

What should marketers do with the savings?

For brands, the corresponding ambition is creating something compelling enough to change what people expect. Wonder depends on encountering something we couldn't have specified in advance and recognizing its value. An audience can trust a brand's judgment while remaining unable to anticipate what it does next — that combination may keep the relationship alive.

An account of what people currently prefer can only do part of the job. Marketers must also exercise judgment about what might interest people once it exists. Cheaper preliminary research should fund exactly that: work with unfamiliar audiences, limited tests of ideas whose value remains uncertain. The organization gains evidence about possibilities it might otherwise have discarded.

A reliable prediction can inform a choice without resolving it. As the earlier essay Peak Analytical Humanity argued, growing analytical capability places more weight on the choices we make with it. The forecast contributes to judgment; the ambition remains ours. Until AI can anticipate every meaningful change in human preference, we still have to put ideas in front of people and see what they make of them.