Where ideas meet the crowd. Understand what is likely to happen, why people react the way they do, and how to respond.
Synthetic user research, grounded in behavioural science. Test questions, ideas, concepts, and messaging against AI personas.
Who's it for?
Built for teams that need a faster read on likely reactions before research becomes slow, expensive, political, or simply too cumbersome to run every week.
Mass gives you a fast way to test ideas against hundreds of personas in minutes, so you can explore likely objections, sentiment shifts, trade-offs, and second-order effects before committing serious budget or political capital.
- Research teams can model diverse reactions without recruitment delays, giving them a quicker way to pressure-test concepts before formal studies begin.
- Universities can explore hypotheses before expensive data collection, helping shape better studies before time and grant money disappear into the floorboards. How universities use Mass.
- Consultancies move faster without letting research swallow the budget, which is useful when clients want answers on Thursday and methodological purity on Friday. Some white-label our outputs. Some hand us the hard part.
- Governments can test policy messaging, likely public reaction, and competing scenarios before full consultation, instead of discovering obvious problems once the press release is already on its way out. How governments use Mass.
Use Mass for concept, message, policy, and scenario exploration; research design and wording; and finding objections and splits early. Do not use it for market sizing, conjoint, representative sampling, statutory consultation, or standing in for lived experience.
No recruiting. No scheduling. Fewer costly detours dressed up as process.
Features
Hundreds of grounded personas to forecast behaviour, demand, and reaction, with explainable outputs for research, policy, and strategy work.
Independent Personalities
Run hundreds of personas calibrated to real behaviour patterns, each with distinct traits, context, and memory. Model reactions across age, location, personality, and circumstance instead of relying on a single synthetic everyperson who somehow speaks for everyone.
Multi-Model Technology
Mass combines leading models to generate behavioural forecasts. Personas reason from distinct perspectives, and cross-model checks help keep outputs robust when the question is politically sensitive, commercially messy, or too important for one model's hunch.
Behavioural Forecasting & Analysis
Get sentiment, reasoning, themes, segmentation, and recommendations with confidence scoring. Reports show not just what people may do, but why, where disagreement is strongest, and which groups are likely to diverge. The numbers describe a simulated cohort, not a measured sample.
Fast Results
Get results in minutes, not months. Test scenarios, messages, and demand signals before you commit to slower research, procurement cycles, or consultation processes that are expensive to reverse once they start moving.
Custom Audiences
Target a specific segment or run a broad spread of perspectives across age, location, personality, and behaviour. Useful when you need to compare likely responses across audiences rather than average them into mush.
Research-Grade Quality
Personas are validated for internal consistency and updated with real-world events, geographical trends, and cultural shifts, so the simulation is not frozen in a cultural moment that ended six months ago.
Data Grounding
Bring the world: surveys, consultation responses, customer feedback, briefs, policy, news, and stimulus such as ads, letters, or concepts. Mass analyses that material so personas react to your context rather than a generic prior. The forecast from that world is directional: it does not grow your sample or add statistical confidence. The report says what it was grounded in. On the platform you can attach images and include research in the prompt; consultancy can ground a run in fuller datasets.
API Access
Integrate Mass into your workflows with API access for automated research pipelines, custom integrations, and internal tools, whether you are supporting an insight team, a research unit, or a policy workflow.
All these capabilities are available in the app. If you want to use the platform in the browser, you can do so by signing up for an account. The platform includes multi-model checks. Grounding and bring your own model are coming in September.
How it works.
From context to forecast in three steps. The point is not to pretend a simulation is the final word, but to get a faster, more inspectable first pass at likely reactions.
1. Build the world - Add the context: briefs, policy, news, financial signals, or your own research inputs. Choose the demographics and behavioural profiles you need so the simulation starts from your world rather than ours.
2. Run the personas - Hundreds of personas evaluate your prompt from distinct perspectives, with cross-model checks to keep outputs robust. They debate, respond, and diverge in ways that make disagreement visible rather than smoothing it away.
3. Use the forecasts - Review sentiment, reasoning, themes, segmentation, and recommendations with explainability and confidence scoring. Use the outputs to refine strategy, sharpen research design, or decide what deserves real-world testing next.
Every step is also available as a service. Our consultancy can build the world, run the personas, and write the forecast with you, or for you, and we will show our workings either way.
Meet some of our personas
Our personas cover a wide range of backgrounds and behavioural profiles, giving you a broader view of how different people may react.
Example analysis
Mass runs hundreds of personas across leading models to produce behavioural forecasts and scenario analysis. The outputs are designed to be read, challenged, and compared, not just exported and admired like a suspiciously glossy deck.
What do you think about remote work becoming the new normal? [cohort/Tech Enthusiasts in the United Kingdom]
How do you think AI will impact your daily life and work in the next 5 years? [cohort/Tech Enthusiasts in the United Kingdom]
What changes would you make to live more sustainably? [cohort/Environmentally Conscious in the United Kingdom]
What features are most important to you when buying a new smartphone? [cohort/Urban Explorers in the United Kingdom]
What role should governments play in regulating social media platforms? [cohort/Health Conscious in the United Kingdom]
What do you value most in a vacation destination? [cohort/Urban Explorers in the United Kingdom]
What kind of advertisements most influence your purchasing decisions? [cohort/Tech Enthusiasts in the United Kingdom]
How important is sustainable packaging when shopping for groceries? [cohort/Health Conscious in the United Kingdom]
What is your biggest challenge in maintaining a healthy lifestyle? [cohort/Health Conscious in the United Kingdom]
App, platform, or consultancy.
Mass is more than just an app. It's a platform for teams that want to run the work on their own computer, and a consultancy for teams that need help with the research, design, and interpretation of the work. The consultancy is Synthetic Research-as-a-Service (SRaaS).
The app is for teams that want to run the work on this computer: private, local, yours.
The platform is for teams that want an account in the browser: inspectable reports at scale, against their own context.
The consultancy is SRaaS. It is for questions that do not fit a prompt. We design the research, build the audiences, run the simulation, and interpret the results with you.
Both routes include custom audiences, dedicated environments, API access, and privacy-sensitive deployments. Reduced pricing is available for non-profits, academic institutions, and public-interest work.
Frequently asked questions
Insights
Insights, research, and updates from the Mass team.
New Models, Same Hive
Aug 7, 2026
We have quietly re-housed the swarm. Mass now runs on the latest generation from Google, with a sharper Pro tier for the heavy lifting. For certain clients, we also route work through the newest Chinese models, because when you care about prediction, you go where the calibration is.
Features
Dead Eyes, Melted Tools, and the One Thing You Cannot Fake
Dec 21, 2025
We stress-tested 14 image models for persona profile photos. The results were not "roughly the same". They were different species.
Research
Same Persona, Seven Mirrors: Persona Validation
Dec 17, 2025
We've conducted a comprehensive validation study of our AI personas to ensure they are realistic, coherent, and useful for market research. This study uses the latest LLMmodels to assess the internal consistency and realism of our personas, ensuring they represent believable, coherent human beings.
Research
The Reality Gap: Decoding the 2025 Christmas Ad Season
Dec 15, 2025
Analysing how specific slices of British society respond to Christmas adverts, using Mass Vision's persona-based sentiment analysis engine.
Research
The Science Behind Mass
Oct 24, 2025
The evolution of market research is witnessing a paradigm shift from traditional human-centered methodologies to AI-powered persona simulations. Here we examine the scientific foundations supporting the persona approach for Mass and demonstrates why AI personas offer advantages over conventional focus group methodologies for initial market validation and feedback generation.
Research
Meet Our Personas
Oct 20, 2025
Discover our diverse collection of AI personas representing real people from around the world. Each persona brings unique perspectives, backgrounds, and insights to help you understand your audience better.
Features