Understanding how an audience weighs one offering against another — including competitors that do not yet exist — is central to consequential positioning and investment decisions. Artificial Societies lets organisations benchmark brand and product perception against real and hypothetical competitors, by constructing networks of 200 to 3,500 interconnected AI personas and measuring how they compare the options. Because competitors can be described hypothetically, teams can test a rival's likely move, or their own unreleased positioning, before either is real, with results returned within 24 hours.
Competitive perception is difficult to research honestly: asking real participants to compare brands is slow, and it is impossible to survey reactions to a competitor or product that has not launched. Simulation removes both limits. Artificial Societies models how a representative audience trades off attributes, price, trust, and positioning across a set of named and hypothetical competitors — using personas grounded in real-world observations validated against human panels — so an organisation sees where it wins, where it loses, and why.
A network of AI personas is constructed to represent the relevant audience, then presented with the competitive set — the organisation's own offering, real rivals, and hypothetical entrants described by their positioning. Reactions are measured across the network: preference, perceived differentiation, switching drivers, and the segments where each option is strongest. Every result opens to individual-level reasoning, revealing the specific perceptions that drive competitive advantage rather than a single share-of-preference number.
Typical questions include how a brand is perceived relative to incumbents, how a proposed repositioning would change that standing, how audiences would react if a competitor made a specific move, and which differentiators actually shift preference. Testing hypothetical competitors is particularly valuable for war-gaming a market before committing to a strategy.
Yes. Because the audience and the competitive set are simulated, a hypothetical competitor or an unreleased positioning can be described and tested before it exists in the market.
Brand trackers measure current perception over time among reachable audiences. Competitive perception testing lets an organisation model how perception would change under a new positioning or a competitor's move, including for hard-to-reach audiences, with individual-level reasoning behind every result.
Any audience for which diverse observations exist, whether public signals or client first-party data. Artificial Societies is strongest for high-value, hard-to-reach audiences where traditional competitive research is impractical.