Product · Overview

See your decisions
before you live with them.

Imagine All The People builds the populations your decisions will encounter, simulates their reactions, and reveals what could make those decisions succeed, fail, or tip.

See how it works

Population synthétiqueAperçu · 10 000 000 individus

Nadia#01 204 771
Gérard#02 918 043
Inès#03 447 190
Claire#04 812 337
Samuel#05 663 028

10 000 000individus
24structures comportementales
Availableentretiens individuels

Calibrated on public data. No personally identifying data.

A different scale of understanding

A population is not
an average.

10 000 000individus
→
1individu
→
son raisonnement
→
the population
→
1decision

Imagine All The People lets you move continuously from the population to the individual — then from the individual back to the collective structure — to inform a decision before you make it.

A decision · simulated

What happens if
TGV prices become
more predictable ?

DecisionPricing policy
PopulationLong-distance travelers
Scenarios3 compared
HorizonOne week

Predictability outweighs price level, including among the most price-sensitive travelers.

TGV train passing through a French city at the end of the day — The system turns a pricing question into a simulatable problem, interviews the population, and compares possible futures.

Lecture décisionnelleExtrait du dossier

Pricing grid announced three months in advanceBroadest support
Price cap per journeyAdhésion nette
Dynamic pricing unchangedRéserve persistante

Price uncertainty acts as a cost in its own right. Reducing it changes booking behavior even before prices themselves change.

Qualitative output: no quantified projection is produced without dedicated calibration.

Inside the product

From decision
to simulation.

Build

The population is instantiated.

Coherent individuals, not average segments.

Each individual has a situation, constraints, and habits that are mutually consistent, calibrated against public statistics and, where relevant, client data.

Population synthétiqueTypologies structurantes

Inès#03 447 190
Claire#04 812 337
Samuel#05 663 028
Karim#06 105 552
Michèle#07 233 908

Immediate adoptionConstraints acceptable
Conditional adoptionSubject to a trade-off
Wait-and-seeDecision deferred

No personally identifying data enters the system.

Interview

Agents conduct the interview.

Contextual follow-ups, not a closed questionnaire.

The agent identifies a point of friction during the interview and returns to it until the reasoning is explicit.

Entretien · agent dynamiqueIndividu #07 233 908

AgentIf this service were available next month, would you use it?
Individual #07 233 908Probably, provided I knew what it would replace. I already use two tools for the same thing.
AgentWhat would make you give up one of those two tools?
Individual #07 233 908Knowing I could keep my history. Starting from scratch costs me more than the subscription itself.

Follow-up 04/09 · every answer stays attached to the individual who gave it.

Simulate

The scenarios are compared.

The entire population is exposed to each option.

Reactions are replayed across the full population, scenario by scenario, over the selected time horizon.

Comparaison de scénariosHorizon · six mois

Scenario A · Phased rollout, announced in advanceBroadest support
Scenario B · Immediate rollout with compensationAdhésion nette
Scenario C · Immediate rollout with no compensationRéserve persistante

ComparisonScenario by scenario, same horizon

Individual reactions

« Finally, clear and simple. I book without stress. »
Individu #04 812 337
« Interesting, but I want to see the terms first. »
Individu #05 663 028

Qualitative reading drawn from the simulated interviews, without extrapolation.

The journey

See the journey,
from a decision to its recommendation.

From the initial decision to strategic and operational recommendations: follow the process through an illustrative example about raising prices.

No run is launched from this page.

Step 01 / 05

The decision

The user brings a decision to inform and its constraints, not a list of characters or a written questionnaire. The system reframes it as a simulable problem and identifies the populations involved.

Decision under reviewIllustrative example

Question askedShould we raise the price of our offer, and with which adjustments?
ConstraintPreserve the relationship with existing customers.
Options to compareThree, defined by the user.

Option A
Keep the current offer and price.
Option B
Raise the price with the offer unchanged.
Option C
Raise the price while adjusting what the offer contains.

A conceptual presentation of the entry point, not an input screen from the application.

01The decision

The user brings a decision to inform and its constraints, not a list of characters or a written questionnaire. The system reframes it as a simulable problem and identifies the populations involved.

Decision under reviewIllustrative example

Question askedShould we raise the price of our offer, and with which adjustments?
ConstraintPreserve the relationship with existing customers.
Options to compareThree, defined by the user.

Option A
Keep the current offer and price.
Option B
Raise the price with the offer unchanged.
Option C
Raise the price while adjusting what the offer contains.

A conceptual presentation of the entry point, not an input screen from the application.

02The population

The system builds the synthetic individuals concerned by this decision, from the characteristics that make it discriminating.

Population concernedSynthetic individuals

#02 918 043
#04 812 337
#05 663 028
#08 470 216
#09 774 305

Construction axes
Tenure with the offer, intensity of use, price sensitivity, availability of an accessible alternative.
Why this population
These characteristics separate those for whom the increase changes a purchase decision from those for whom it changes nothing.

No personally identifying data enters the system.

03The scenarios

The three options are replayed on the same synthetic individuals. Agents conduct the interview, probe friction points and surface the arguments.

Entretien · agent dynamiqueOption C · follow-up

AgentIf the price went up and the offer included more, would you stay a customer?
Individual #08 470 216It depends what is added. If it is an option I do not use, I am paying more for nothing.
AgentWhat would justify the increase in your eyes?
Individual #08 470 216Saving time on what I already do every week. The rest I can do without.

What changes between A, B and C
The same individuals face all three options: the arguments they give differ, and the breaking points are not in the same place.
What this does not establish
Questioning the same individuals allows options to be compared at constant population; it does not establish causality and does not guarantee identical answers.

Qualitative reading drawn from the simulated interviews, without extrapolation.

04The recommendation

Responses to questions, brainstorming sessions and analyses inform the system’s strategic and operational recommendations. A stated intention in simulation is not a sale.

RecommendationA deliverable distinct from the individual answers

If the scenario comparison showed greater acceptance of the price increase when the offering is adjusted, the recommendation would be to favour that option, subject to its economic feasibility.

The proposed actions would be to specify the adjustments to the offering, explain their value in the launch message and prepare responses to the objections identified.

Remaining uncertainties would guide further questions or analyses before finalising the implementation details.

Qualitative output: no revenue or margin projection is produced here.

05Go deeper or replay

Two distinct possibilities: change an option and compare the reactions of the same population again; or investigate an uncertainty further through new questions, new analyses and, if needed, new populations.

Possible next movesTwo distinct operations

Replay on the same population

The user changes an option and compares the reactions again, keeping exactly the same synthetic individuals.

Investigate further

The system can ask new questions, re-run analyses or generate new populations to resolve an identified uncertainty.

These next moves are possibilities open to the user, not an automatic step in the journey.

A new category

Observe.
Generate.
Simulate.

ObservePanelsMeasure reality.

GenerateIA générativeGenerate possibilities.

SimulateImagine All The PeopleTest what could happen.

ScaleA sample size defined by the research protocolThe number of profiles depends on the tool and setupPopulations that can comprise millions of individuals

DepthQuestionnaires and interviews with real peopleResponses generated from the supplied contextSimulated interviews, follow-up questions and brainstorming

Time dimensionOne-off or repeated measurements over timeTime-based hypotheses explored using the supplied contextSimulated developments over the horizons studied

TraceabilityStudy responses and data, depending on the designSources and history accessible depending on the setupFrom aggregate results to the simulated responses behind them

PanelsObserve

ScaleA sample size defined by the research protocol
DepthQuestionnaires and interviews with real people
Time dimensionOne-off or repeated measurements over time
TraceabilityStudy responses and data, depending on the design

IA générativeGenerate

ScaleThe number of profiles depends on the tool and setup
DepthResponses generated from the supplied context
Time dimensionTime-based hypotheses explored using the supplied context
TraceabilitySources and history accessible depending on the setup

Imagine All The PeopleSimulate

ScalePopulations that can comprise millions of individuals
DepthSimulated interviews, follow-up questions and brainstorming
Time dimensionSimulated developments over the horizons studied
TraceabilityFrom aggregate results to the simulated responses behind them

Under the hood

A simulation infrastructure,
not a collection of prompts.

Coherent population
Individuals whose situations, constraints, and behaviors remain mutually consistent at every scale.
Dynamic agents
Agents that conduct interviews, detect friction, and follow up instead of simply running through a questionnaire.
Temporal simulation
The ability to replay a decision over a chosen horizon, from one week to several years.
Explainable reasoning
Every aggregate can be unfolded down to the individual interviews and arguments that produced it.

10 000 000Simulatable individuals
1 000+Parallel tests
76Markets covered

A public benchmark for synthetic population generation. The protocol compares generation methods by measuring their deviations from reference statistical distributions. View the benchmark and its limitations

Validation

Simulation only has value
if it holds up against reality.

Validation is part of the infrastructure, just like population generation or interview execution. It relies on four operations controlled by the scientific team.

The protocols, comparison datasets, and their limitations are detailed in the Science section. See validation and calibration →

By function

One infrastructure.
Different decisions for different functions.

Executive leadership, strategy, marketing, innovation, research, public affairs, or human resources: each function can use Imagine All The People for different decisions.

Executive leadership
Strategy
Marketing & Brand
Innovation & Transformation
Research & Insights
Public affairs
Human resources

Explore by function

Your turn

Which decision
do you want to test ?

Testing a price increase
Launching a new offer
Comparing two messages
Preparing a brand overhaul

Free trial, no credit card required. Your data remains confidential.

By decision type

By industry

See your decisions
before you live with them.

A population is not
an average.

What happens if
TGV prices become
more predictable ?