Crisis simulated. Readiness proven.

We run a live, AI-scored crisis simulation for your team, built on your own risk material. Your people make real decisions under pressure, every call is timed and graded, and you leave with a report showing precisely where the response breaks.

Book a scoping call See what an exercise involves

One risk-built scenario, two live simulations, one scored review.

4 hrs
per live simulation
8–30
participants across your teams
Same day
scored after-action review
Z Live feed Exercise
Trending: #Eastport 18,402
PN
Priya Nathan @priya_nathan · 2m

Three ambulances at the Eastport site and they are walking everyone out into the car park. My sister works in that building and she is not picking up.

214 1,908 6,733 412K
DK
Dan Kowalski ★ @dk_reports · 7m

I worked Eastport six years. That line was flagged in an internal audit last March and nothing was done. The document exists. They knew. 1/7

Misinformation Unverified
Virality 94
3,104 18,241 42,806 2.4M
Requires response 04:38
MW
Metro Wire @metrowire · now

We run the Eastport story at 15:00 with the audit claim in it. No statement by then and we report that the company declined to comment.

Your account Silent 41 min Every minute is scored

Illustrative interface views

Fictional, non-operational scenarios only. Built for government, defence, emergency services, critical infrastructure, corporate crisis and communications teams.

What you get

One engagement, three deliverables

Two consultants at a table spread with plans and documents, working by lamplight 01

Build your crisis with professional risk consultants

Our risk consultants work with your team to turn your crisis plan, risk register and standard operating procedures into a bespoke scenario. The exercise reflects your sector, sites, stakeholders, dependencies and decision thresholds.

Your team validates the scenario before the first simulation begins.

A room of participants at monitors during a live exercise, lit only by their screens 02

Two live simulations

The first simulation establishes a performance baseline. A facilitated reflection then converts the findings into practical adjustments before the second simulation tests whether those changes improve performance under pressure.

Baseline, facilitated reflection, retest and measured improvement.

A findings session in a darkened boardroom, charts on a lit wall display 03

A scored after-action review

Receive a scored account of what happened, including decision latency, coordination, team and participant performance, the complete decision timeline and prioritised recommendations. Exported to PDF or Excel.

Evidence generated from the simulations, not reconstructed from memory.

Every engagement includes all three

Scope, format and dates are agreed with you at the scoping call, before anything is built. Your organisation is invoiced directly and pays online, so nobody on your team needs an account with us.

Discuss an engagement

Data handling

Your crisis material is never used to train an AI model

A realistic exercise means handing us your risk register, your escalation thresholds and the things your executives would actually say at three in the morning. That material is treated as what it is: some of the most sensitive information your organisation holds.

  • No training, no fine-tuning

    Your scenario material and every decision made during a simulation are processed to run your exercise and nothing else. They are never used to train or fine-tune any model, ours or a vendor’s.

  • Private inference on AWS

    Language models run through Amazon Bedrock. Prompts and responses are not shared with the model providers and are not retained by them. That is an AWS contractual commitment, not a setting we toggle.

  • In-region and isolated

    Your data stays in the AWS region agreed with you at scoping. Each client’s material is kept separate; nothing is pooled across engagements.

  • Yours to remove

    Request a full export or permanent deletion of everything from an engagement at any point, during or after it.

Recorded · Communications lead · T+00:58
Forty minutes into the silence, a communications lead writes the organisation's first public statement. Legal has not cleared a figure. It publishes anyway.

Corporate Crisis · recorded, not mocked up

What it looks like from inside the room

Your people are issued a working environment of two social platforms, an inbox, a news cycle and a team chat. Then it turns on your organisation. Misinformation moves faster than your approval chain, and every decision is timed and scored while they make it.

  • Posts, messages and coverage arrive live while a window is open. Nothing is pre-baked into a slide.
  • Hostile accounts respond to what your team actually says, not to a script written in advance.
  • Every public statement is checked against the scenario's own record of what is true.
See all seven acts of a full exercise Ninety minutes, act by act
The feed A claim your team could disprove in two minutes, at a quarter of a million views.
A simulated inbox showing an urgent request for comment from a reporter, with three numbered questions and a 6pm deadline.
The inbox A reporter with three specific questions, publishing at 6pm whether you reply or not.
A participant reporting a post as misinformation, having typed in the verified facts that contradict the claim.
The counter-move Reporting the lie is scored on whether the cited fact actually disproves it.

Recordings of a delivered exercise · scenario content is simulated

Key highlights

Four things that make it behave like a real crisis

These are the design decisions that separate an exercise here from a scripted tabletop, and the reason your team cannot settle into a comfortable answer.

01

Multi-agency coordination

Crises do not respect a single jurisdiction, and operational success usually depends on coordinated action across several. Every participating agency works inside one shared environment, entering decisions concurrently rather than waiting its turn. The pressure lands where multi-agency response actually fails: asset deployment across organisations, communication workflows, and joint command.

02

Information asymmetry

No agency ever holds a complete assessment of the theatre, so nobody is given one. Information is deliberately partitioned into role-specific intelligence streams, stripping out the artificial clarity that makes conventional exercises easy. What is left is the real bottleneck: whether teams can build the protocols to validate each other’s information quickly enough to construct a shared operational picture.

03

Real-time adaptive injects

There is no storyboard to work through. A condition-dependent engine evaluates what participants actually do and branches the scenario in real time, so the incident escalates or settles according to the speed and quality of the response. That removes predictability entirely. The trajectory follows the decisions, not a script.

04

Team interdependency and storyboard synthesis

Operational domains are interdependent, so no decision is made in isolation. One team’s intervention alters the conditions every other team is working in, because a central synthesis layer feeds the output of each module straight back in as live input to the others. Separate operational threads run as a single engine, and teams have to absorb the compounding friction their counterparts create.

Inside an exercise

What your people are actually doing

Not discussing a scenario around a table. Working a live picture, defending public statements, and watching a conversation they do not control.

Hands working a tactical map marked with incident points, cordon rings and movement arrows

Command works the picture

Positions, cordons and priorities change as the incident develops, and each team only sees the part of it they have been told.

Three people at laptops in a darkened room drafting public responses under time pressure

Comms defends every statement

Anything your team says publicly is checked against what the scenario knows to be true, and challenged when it does not hold.

A column of abstract social posts multiplying outward, some glowing as they gain traction

The public conversation runs regardless

A simulated feed, inbox and news cycle carries on with or without you, and misinformation spreads faster than corrections.

Illustrations, not screenshots

Exercise types

Three ways to put your team under pressure

Same scoring and the same report in every case. We recommend a format once we understand what you are trying to prove.

An emergency operations centre, figures at a wall of maps marked with incidents

Type 01

Multi-agency coordination

Tests the handoffs between your own functions, and with the external agencies you would actually depend on. Each team sees a different slice of the truth.

  • Deliberate information asymmetry between teams
  • Role-scoped channels, resource requests and negotiation
  • Best suited to command groups and executive teams
Responders in high-visibility jackets directing a crowd away from a venue at dusk

Type 02

Site and public safety response

For organisations responsible for physical sites and the people in them. A live map of a real location, with a crowd simulation running underneath that reacts to every decision.

  • Evacuation and cordon decisions with modelled consequences
  • Casualty condition deteriorates while decisions are pending
  • Built on real building footprints and street layouts
A participant's simulated desktop with the crisis response team chat open beside a social feed

Type 03

Corporate Crisis

For communications, legal and leadership. A simulated feed, inbox and news cycle turns on your organisation while misinformation moves faster than your approval chain.

  • Simulated posts, direct messages and news coverage
  • Every public statement is fact-checked against the scenario
  • Hostile actors that respond to what you actually say
Walk through a full exercise

Who this is for

Built for organisations that cannot afford a first attempt

Government & defence

Inter-agency coordination and public information under political and media pressure.

Typical exercise

Coordinated attack during a public event, with competing agency priorities.

Emergency services

Mass-casualty response, cordon and evacuation decisions, multi-service command.

Typical exercise

Explosion at a crowded venue with a suspected secondary device.

Critical infrastructure & utilities

Service loss, cascading failure and communicating with a public that needs answers.

Typical exercise

Extended outage affecting hospitals and transport, cause unknown for hours.

Corporate crisis & communications

Product harm, data compromise or executive conduct playing out in public in real time.

Typical exercise

A safety allegation goes viral before your own investigation has confirmed anything.

Media & public affairs

Verification pressure, misinformation, and reporting responsibly under uncertainty.

Typical exercise

Fabricated footage of a live incident spreading faster than official confirmation.

Not on this list?

The scenario is built from your material, so the sector matters less than the risk. If you can describe what keeps you awake, we can build an exercise around it.

Tell us your risk

Safety

Safe to rehearse, and safe to hand to us

Fictional and non-operational

Scenarios are invented. They are built to resemble your exposure closely enough to be uncomfortable, without documenting a real target, a real vulnerability or a real operational plan.

Nothing touches a real platform

The feeds, inboxes, news outlets and messaging your people use during the exercise are all simulated inside the environment. Nothing a participant posts can reach the public internet or a real system.

Settled before anything is built

Exactly what material you share, and what is retained once the exercise is over, is agreed at scoping rather than assumed. Bring your security team to that call.

Questions

Before you take this to your leadership

How is this different from the tabletop exercise we already run?

A tabletop is usually a structured discussion following a fixed script, closed out by a subjective debrief. Here the situation advances on a clock and reacts to what your team does or fails to do, so late decisions have consequences you can see. The debrief is then assembled from measured decisions rather than from what people remember agreeing.

Who designs and runs it?

A Prophyion risk consultant. They build the scenario with you from your own material, facilitate both simulations, and present the scored after-action review. You are not being handed software to operate yourself.

Do our people need training on the system first?

No. Participants receive a role briefing and work in a browser. The interface is deliberately plain, because the exercise is testing judgement and coordination, not anyone’s proficiency with a tool.

Can you build it around our existing crisis plan?

That is the preferred input. The exercise is most valuable when it tests the plan you would genuinely follow, including the assumptions inside it that have never been put under load.

How long until the first simulation?

Most of the elapsed time is your review, not our build. Once we have your material a draft scenario comes back quickly, and we then iterate until you are satisfied it reflects your real exposure. We will give you a firm date at the scoping call.

What exactly do we hold at the end of it?

A scored after-action review covering both simulations: a narrative of what happened, metrics per team and participant, the complete decision timeline, and prioritised recommendations. Exported to PDF or Excel for a board or audit pack.

What does an engagement cost?

We go through it at the scoping call, once we know which format you need and how many people are taking part. An engagement is quoted as a single figure covering the scenario, both live simulations and the scored after-action review, so there is nothing to reconcile afterwards.

Get started

Start with a scoping call

Tell us what you are actually worried about. We will come back with a scenario outline, a recommended format, and what an exercise would look like for your team.

  • A person replies, not an automated sequence.
  • No commitment, and nothing gets built until you approve the scenario.
  • Bring your security or procurement lead if it helps.

Prefer to circulate something internally first? Say so in the form and we will send a short summary you can forward.

We use your details only to respond to this enquiry.

Every agency can rehearse chaos, safely