Beta · one free run

Rehearse it with the market before you commit.

Describe a product, a price, a campaign or a proposal. A simulated audience built from Nextatlas behavioural data answers it: who takes it up, who refuses, and the objection that keeps coming back. You reach the expensive research already knowing which questions are worth paying for. The audience is simulated rather than surveyed, and every result shows its work.

One free run · No credit card · Your briefs stay private
proximalab.ai/simulator/sim-2025-de-pbc
Example · a product
Plant-based chocolate · Germany · Gen Z
Simulated adoption
55%
of the population
Sentiment
+0.42
weighted
Quotes
20
profiles answered
Adoption across roundsscenario
Why it exists

Know the objection before the room does.

Trend data tells you what is already happening. Proxima tells you how an audience would answer something that has not happened yet. Describe it, and a simulated audience drawn from Nextatlas behavioural data responds: a model decides take-up across every profile, and a sample of them answers in words, with the reasoning attached. Throw out the weak versions in an afternoon and spend the research budget on the questions that survive.

What you can decide

Four decisions you can pressure-test today.

Each card starts from a real brief and names the decision it serves. It also names where the answer stops: knowing that before you run is worth more than discovering it halfway through a presentation.

Product

Should we launch at this price?

An oat-based barista milk for speciality coffee bars in Italy, at €2.90 per litre.

Scenario Simulator. You learn who takes it up, how far it travels, and the objection that keeps recurring.
Where it stops: Adoption is computed by a rule-based model across the population. A separate sample provides the written reactions. It is not a sales forecast.
Service

Will they commit to a subscription?

A monthly bike-repair subscription for commuters in Milan: unlimited fixes for €12 a month.

Scenario Simulator. You learn who signs up and what they hesitate over: price, trust or effort.
Where it stops: Profiles decide once, as of today. Churn, renewal and usage over time are not modelled.
Campaign

Which version earns the budget?

Two versions of a message for a phone trade-in scheme: “Repair, don't replace” against “Your old phone is worth more than you think”.

Campaign Pre-Test. You learn how each version lands by segment, whether it gets shared, and the themes running through the replies.
Where it stops: The panel is also asked whether it would buy because of the message. Ignore that column when nothing is sold.
Proposal

How will people take this?

A four-day week pilot for a 200-person design studio: same pay, Fridays off, core hours Monday to Thursday.

Scenario Simulator. You learn who welcomes it, who resists, and the reasoning on both sides.
Where it stops: Profiles read it as something new entering their world and answer adopt or not. Support, voting and compliance are not modelled.
Four modules

Four ways to interrogate the same brief.

01 · Scenario Simulator
Will they take it up?

Put a product, a service or a proposal to an audience built from behavioural data. A diffusion model decides take-up across every profile; a sample of them also answers in words. You get the curve, the segments that split, the objections and a written briefing. The free run is this module, at 100 profiles.

See a full report
02 · Living Personas
Put it to nine standing archetypes.

Nine editorial archetypes written by hand rather than mined from the data, plus any persona you define yourself. Interview them one at a time or run them as a focus group, and watch the trends each one follows move. Full accounts.

Book a demo
03 · Campaign Pre-Test
Which version earns the budget?

Enter two or more versions. A panel reacts to each one: sentiment by segment, whether they would share it, whether they would buy because of it, and the themes running through the replies. When the panel cannot separate them, it says so. Full accounts.

Book a demo
04 · Category Engine
What is moving around the theme?

Seed a theme with up to three terms and get the trends around it with their measured growth, refreshed when you ask. A what-if adds a qualitative reading of how the map could shift, with no simulation and no magnitudes. Full accounts.

Book a demo
The deliverable

Written like an editorial. Read like a report.

Every run produces a briefing you can take into a meeting without translating it first. If the briefing fails to generate, the page says so, and the numbers and quotes are still there.

An adoption curve, labelled as modelled rather than measured
Objections in the words the sampled profiles used
Recommendations tied to the evidence behind them
The whole run exportable as JSON
Example report · a product
Plant-based chocolate
500 profiles · 20 rounds · one recorded seed

A premium plant-based chocolate at €4.99 finds a strong, vocal protagonist cohort of Gen Z urban ecoconscious adopters, and converts 55% of the target population across the simulated rounds…

Simulated adoption
55%
Sentiment
+0.42
Profiles asked
20
One seed · fully reproducible
Proof

Every number shows its work.

Open the box
Every run records where its audience came from, which Nextatlas dimensions came back real, which fell back to defaults, and how many profiles answered. No figure arrives without its origin attached.
Run it again
A recorded seed replays the same audience and the same curve, given the same population inputs, the same configuration and the same engine version. The written answers are generated fresh each time. Change one declared parameter and Compare confirms it is the only difference.
Read the words
Adoption is a number the model computes over every profile. Objections are sentences, and each one comes from the sampled profile that was asked: a simulated voice, not a person's.
One free run
Every confirmed account gets one Scenario Simulator run, at 100 profiles, with no card and no contract. The other modules are for full accounts while the beta is small. Ask us for one.
How it works

From a brief to a briefing.

STEP 01

Write the brief

What you are considering, who it is for, where and when. A few sentences is enough, and nothing is invented for a detail you leave out. Then confirm the population reference: the theme Nextatlas uses to build the audience.

STEP 02

The audience answers

Proxima assembles an audience from Nextatlas behavioural data. A diffusion model decides take-up across every profile, round by round, and a sample of profiles is asked to answer in words: that sample is where the quotes and objections come from. The profiles are built from behavioural data, not real respondents.

STEP 03

Read the argument

A briefing keeps what the model computed apart from what the sampled profiles said, turns the evidence into hypotheses, and states what the run does not settle. Change one detail, run it again and compare the two.

1,000
Profiles in one audience
on full accounts; the free run uses 100
3
Behavioural dimensions from Nextatlas
demographics, professions, attitudes
40
Rounds of diffusion per run
on full accounts; simulated rounds, not months
1
Seed replays the audience and the curve
same inputs, configuration and engine version; the written answers vary
Pricing

The beta is free.
Pricing comes after validation.

Available now
Beta
Everything a confirmed account gets today. There is no paid plan yet, and we would rather say so than invent one.
Freeduring the beta
One Scenario Simulator run, at 100 profiles
The other modules on full accounts. Ask us for one.
Your runs stay private to your account
No contract and no card
Later
Pricing follows validation, not the other way round.
Not priced
Seats, quotas and support tiers are undecided
We would rather learn what the method is worth first
Talk to us if you want to shape it
Nextatlas is an EU company and its signal layer is aggregated and anonymised at source. Proxima itself runs on Railway and Vercel, currently in a US region. If data residency matters to you, ask us before you put anything sensitive in a brief.
FAQ

Questions strategists ask us.

Anything a person could react to: a product, a service or subscription, a campaign message, a concrete proposal. Describe what it is, who it is for, the context and the question you have. The Scenario Simulator asks whether the proposal would be taken up, so it fits best when taking it up is the question. To compare versions of a message, use the Campaign Pre-Test. It is not suited to usage over time, renewals, votes or compliance, none of which are modelled.
Ready when you are

See the reaction
before you place the bet.

One free run, no card and no contract. Read the limits on the result, then take the questions it raises to real people.