Real data
Geography, demographics, income and spending, category interests, and behavior come from proprietary surveys and government data. Consumers carry that buying power into the run.
Simulate the marketGive your data a brain.
SynthEngine combines survey and government data into an interconnected ontology of a synthetic population, then layers AI reasoning onto it to simulate how people think, interact, and make decisions. Test what could influence them to choose your product, switch to a competitor, or change their minds. Explore the forces behind human decisions before you spend on the real world.
Decisions are made before the market responds.
A price change, a launch, a campaign, a rival’s entry: the commitment is made, and only then does the market answer. By the time a new round of research comes back, the money is already in motion. What people tell each other, and what a new law changes about what they can buy, rarely land in the same picture.
- The commitment is made first.
- The response arrives later, and in pieces.
- Word of mouth and outside events rarely share a model.
How it works
A scenario runs on a population that was built from evidence, then the run stays open to inspect.
-
01
Build the population from real data
Proprietary survey data and government data supply geography, demographics, income and spending, category interests, and behavior patterns. The synthetic population is built from those records.
-
02
Introduce the scenario
A product launch, a pricing change, a position, a competitor’s entry, or a campaign is placed in front of that population.
-
03
Agents interact, and the world evolves
Consumers respond and talk. Word of mouth spreads across locations. The simulated world can update with events such as a political change or a new law.
-
04
Analyze the outcomes
Follow how consumers in the run responded, where talk traveled, and which constraints held. The path from the source data to the outcome stays open to inspect.
Under the hood, LLM reasoning, Jev (a fast System 1 judgment layer), RICH/OASIS social simulation, and the FAST quantitative choice engine work as one traceable system.
The population has a place.
Consumers are placed from the geography in the survey and government data. A scenario is introduced there. Talk moves across locations, and the world can change when events do.
Drag to look along the street. Shops, traffic, and a town-hall rally, looping.
Loading the population globe…
What you can explore
-
Pricing
Introduce a price change and follow who stays with the brand, who switches, and where word of mouth travels.
-
Product launch
Place a new product in front of consumers who already have budgets, habits, and a location.
-
Competitor entry
Explore how an incumbent’s buyers may respond when a rival enters the same market.
-
Campaigns
Study a message before media spend, including how it spreads from person to person.
-
Positioning
Compare ways of framing the same offer against the same population.
Why SynthEngine
Social dynamics
Agents interact. Word of mouth spreads across locations, as part of the run.
A live world
Events such as a political change or a new law can enter the simulation, so the market is not frozen on the day the file was exported.
Traceable
Reasoning, social interaction, and the quantitative choice stay linked, so an outcome can be followed back to the evidence and the scenario.
Questions
What is SynthEngine?
SynthEngine is a market intelligence and counterfactual simulation platform designed to explore how markets, consumers, and competitive environments may respond to strategic interventions before capital and resources are committed. It fuses heterogeneous data into an interconnected market ontology, then applies a synthetic cognitive layer to model consumer decision-making, behavioural heterogeneity, competitive dynamics, and potential market outcomes. From pricing and product launches to political analysis and policy scenarios, SynthEngine transforms fragmented evidence into an inspectable environment for strategic experimentation.
Where does the population come from?
SynthEngine constructs synthetic populations grounded in government microdata, household consumption surveys, socioeconomic indicators, and other relevant empirical datasets. These inputs establish the population's demographic composition, geographic distribution, household economics, and supported behavioural characteristics. Proprietary surveys and category-specific research can enrich the underlying evidence where available. Each variable retains its epistemic provenance, distinguishing directly observed attributes from derived estimates, modelled probabilities, and synthetic assumptions, so that population realism is not conflated with behavioural certainty.
What can a team explore?
SynthEngine enables structured scenario exploration across pricing, product launches, market positioning, competitor entry, promotional strategies, distribution changes, and campaign interventions. Teams can examine alternative configurations against a consistent synthetic population, isolate individual variables, and compare conditional outcomes. The same framework can be extended to political analysis, public sentiment, policy interventions, and societal scenarios where the necessary evidence and behavioural models are available.
Does SynthEngine replace a survey?
No. SynthEngine complements primary research, consumer surveys, and market intelligence rather than replacing them. Empirical datasets establish the population and constrain the model, while synthetic simulation provides a repeatable environment for exploring hypotheses, interrogating assumptions, and comparing counterfactual scenarios. Its purpose is to extend the analytical value of existing research before decisions advance to more expensive validation, fieldwork, or market execution.
Does it tell you exactly what the market will do?
SynthEngine does not claim deterministic certainty about human behaviour or future market outcomes. It generates conditional outcomes under explicitly defined scenarios, population characteristics, decision rules, and model assumptions. These outputs help teams examine potential demand shifts, substitution effects, competitive responses, and behavioural constraints. Their predictive reliability depends on empirical calibration, the quality of available evidence, and independent validation against observed outcomes.
What happens during a run?
A scenario is translated into a structured intervention within the market environment and evaluated against the relevant synthetic population. The quantitative engine adjudicates consumer choices and calculates supported market outcomes, while enabled cognitive and social components can represent selected interactions, information exposure, and behavioural responses. The broader world model is designed to incorporate temporally relevant events, such as political developments, regulatory changes, or new legislation, where the necessary event representations and integrations are available. Scenario definitions, model assumptions, decision pathways, and generated outcomes are intended to remain inspectable, allowing users to examine not only what changed, but the mechanisms and constraints behind the result.
What runs under the hood?
SynthEngine's architecture combines empirical population calibration, a structured market ontology, an interpretable quantitative choice engine, bounded large language model reasoning, and specialised cognitive and social simulation components. FAST serves as the quantitative decision authority for supported market simulations; Jev represents a System 1 style rapid-judgment layer, while RICH and OASIS provide components for richer cognitive and social interaction experiments. The ontology organises entities, attributes, relationships, and contextual events, while provenance and experiment records support traceability. These components have different execution roles and levels of integration; not every component is active in every run, and behaviour generated by uncalibrated components remains a model assumption rather than empirical evidence.
Who is the design partner programme for?
The design partner programme is intended for FMCG companies, brand strategists, consumer insights teams, market researchers, and organisations seeking to investigate complex market decisions using synthetic populations and scenario-based experimentation. Partners bring a category they understand, a commercially or strategically meaningful question, and relevant evidence where available. Together, they can evaluate the applicability of SynthEngine's modelling framework, identify evidence gaps, and establish validation requirements for use cases such as pricing, product launches, competitive disruption, campaign strategy, political analysis, and policy assessment.
How do I join the waitlist?
Submit your work email through the waitlist form to register your interest in SynthEngine and receive communications about product development, design partner opportunities, and access availability. Your email address is intended to be used for SynthEngine-related communication, in accordance with the website's privacy policy.
Design partners
For FMCG, brand, and insights teams. Bring a category you already know and explore a pricing, launch, competitor, campaign, or positioning question with us.