Rehearse the decision before you make it.

Marden

A fictional company.

270 people in European technology and business services. Every figure and sentence on this page is precomputed for it.

Reading the numbers

A range, not a point.

Each headline figure spans the middle half of the modeled outcomes, 25th to 75th percentile. A trajectory shows how the change is expected to unfold over 26 weeks.

Subject

Roles, never people.

Demand load and delivery trajectory are read per department under the plan. The model projects how work is allocated.

Changeprint

A pattern, not a score.

The whole picture of one WhatIF: four headline ranges, an executive read, what changed and where load traveled.

Calibration

4,500+ studies behind the model.

Its response parameters draw on published research into how people responded to organizational change. That sets the bounds; it does not turn any single run into a forecast.

Security

Your data stays yours.

Every simulation runs offline, on an engine we built, and no outside AI ever reads what you share. It never leaves the EU, stays encrypted in transit and at rest, and each client can open only its own results.

Marden

A fictional company.

270 people in European technology and business services. Every figure and sentence on this page is precomputed for it.

Reading the numbers

A range, not a point.

Each headline figure spans the middle half of the modeled outcomes, 25th to 75th percentile. A trajectory shows how the change is expected to unfold over 26 weeks.

Subject

Roles, never people.

Demand load and delivery trajectory are read per department under the plan. The model projects how work is allocated.

Changeprint

A pattern, not a score.

The whole picture of one WhatIF: four headline ranges, an executive read, what changed and where load traveled.

Calibration

4,500+ studies behind the model.

Its response parameters draw on published research into how people responded to organizational change. That sets the bounds; it does not turn any single run into a forecast.

Security

Your data stays yours.

Every simulation runs offline, on an engine we built, and no outside AI ever reads what you share. It never leaves the EU, stays encrypted in transit and at rest, and each client can open only its own results.

Engine

200+ calibrated coefficients.

In a full engagement they span the whole state model: structure, dependencies between departments and how concentrated they are, demand load on roles, delivery effects, and how all of it moves over time. Every run records its version and scenario.

Loadpath

A route, not a verdict.

Where added coordination, dependency or transition load travels between departments over the simulated window. One to three are shown, most consequential first.

Depth

1,500 runs per decision.

The current level for a full engagement, and what makes every output a distribution of outcomes.

Switchpoint

A position in the network.

A role that links functions otherwise weakly connected. It follows from the structure of this Base State and says nothing about the person in it.

Purpose-built

Made for one job: simulating how organizations behave.

Chat-based AI writes one likely word after another, so a question asked twice can get two different answers. Ours has no language model inside; it computes how roles, teams and departments respond to a decision. It is deterministic: pick the same tiles and the figures never change.

Engine

200+ calibrated coefficients.

In a full engagement they span the whole state model: structure, dependencies between departments and how concentrated they are, demand load on roles, delivery effects, and how all of it moves over time. Every run records its version and scenario.

Loadpath

A route, not a verdict.

Where added coordination, dependency or transition load travels between departments over the simulated window. One to three are shown, most consequential first.

Depth

1,500 runs per decision.

The current level for a full engagement, and what makes every output a distribution of outcomes.

Switchpoint

A position in the network.

A role that links functions otherwise weakly connected. It follows from the structure of this Base State and says nothing about the person in it.

Purpose-built

Made for one job: simulating how organizations behave.

Chat-based AI writes one likely word after another, so a question asked twice can get two different answers. Ours has no language model inside; it computes how roles, teams and departments respond to a decision. It is deterministic: pick the same tiles and the figures never change.