FCA-regulated firms must prove they can survive a crisis. Can yours?

For CISOs and Heads of Operational Resilience at UK insurers, banks, and FinTech companies: we're building the Crisis Room Simulator to surface where your crisis response is likely to break and generate the evidence you need to fix it, while there's still time before the regulator tests it for real.

May 2026

PRA DyGIST live stress test launches. A 3-week crisis simulation covering 80%+ of PRA-regulated general insurance market. Scenarios not shared in advance.

March 2027

PS26/2 incident reporting enforcement begins. Firms must detect, escalate, and report operational incidents through a single portal shared with the PRA and Bank of England.

Now

FCA 2026 observations name AWS, Azure, and Cloudflare outages alongside Jaguar, M&S, and Co-op cyber attacks as "severe but plausible" scenarios firms should be testing. Are you?

The board question you need to answer.

The question you can't answer

The regulator asks: “Show us the evidence that your crisis response actually works.”

You filed your impact tolerances on time. Your Important Business Services register is mapped. But you've never actually tested what happens when someone pulls the trigger on isolating a compromised system at 2am. Who calls who? How long before critical services go dark? Does anyone even know the FCA notification threshold's been crossed?

The FCA already knows this is where firms are weakest. Their 2026 observations are clear: some firms “state there's no scenario they wouldn't be able to recover from, but don't include evidence.”

The test that's coming

For insurers, the PRA's DyGIST exercise is a 3-week live crisis simulation unlike anything the sector has faced before. Scenarios not shared in advance. No scripts. No best behaviour. For all FCA-regulated firms, PS26/2 enforcement starts March 2027.

Your annual tabletop (two hours, everyone performing, £50k+ to a consultancy) doesn't come close to replicating this. And the FCA's direction of travel is clear: firms need to move beyond paper-based assurance toward more sophisticated, evidence-based testing.

You're the one who'll be answering the board's questions when it goes wrong. Right now, you're answering with assumptions.

204
Nationally significant cyber incidents in 2025, up from 89
NCSC Annual Review
47
Firms sampled by FCA, with areas of poor practice found in tolerance-setting and scenario testing
FCA Insurance Insights
40%+
Of cyber incidents involved a third party
FCA 2026 Observations
80%+
Of PRA-regulated GI market covered by DyGIST
Bank of England

Designed to show you where your crisis response breaks.

We build a twin of your organisation. Then we break it.

Not your infrastructure. Your people. AI personas of your CISO, CEO, CRO, and functional heads make decisions in real time. You watch the cascade unfold: which handoffs get missed, which tolerances breach, which downstream services nobody thought about until they went dark.

The goal: surface the likely failure points in your crisis response and generate audit-ready evidence. So you can fix it before it matters.

For the CISO

Walk into the next board meeting with evidence, not assumptions. Show the board a documented simulation that either proves your response works or pinpoints exactly what needs fixing. Before the regulator does it for you.

For the Head of Op Res

Your tolerance mapping finally gets tested against something that moves. Monthly, not annually. Every run produces audit-ready documentation automatically, so you can stop manually writing up tabletop findings.

For the Board / CRO

See a clear before-and-after: which gaps existed, which were closed, and the documented evidence trail the PRA expects. The FCA's 2026 observations leave no room for doubt: firms need empirical testing, not judgement-based assurance.

What good looks like: before and after.

Without Crisis Room Simulator

  • Annual tabletop exercise: 2 hours, scripted, everyone on best behaviour
  • £50k+ to a consultancy for a single scenario
  • Paper-based findings report delivered 3 weeks later
  • No evidence your response has actually improved
  • The regulator's stress test arrives and your team faces a live crisis simulation for the first time

With Crisis Room Simulator

  • Monthly simulations: realistic, unscripted, pressure-tested
  • AI twins reveal the real gaps, not the polished version
  • Auto-generated evidence packs, board-ready in minutes
  • Track improvement over time with comparable results
  • The regulator's stress test arrives and your team has already rehearsed the format

Everything you need to prove resilience.

See how your team actually responds under pressure

AI twins of your C-suite and functional heads make decisions in real time. You see who escalates late, where accountability gets fuzzy, and which handoffs fall through. The things a scripted tabletop will never surface.

Know what breaks before it breaks

When you isolate a system, we show you everything downstream that fails: SLAs, customer journeys, third-party dependencies, regulatory notification thresholds. You get the view the regulator wants before they ask for it.

Catch tolerance breaches in real time, not in hindsight

Real-time alerts when simulated response times blow past your stated tolerances. A flag in the moment, tied to the specific service and the specific decision that caused the breach. Not a retrospective finding three weeks later.

Rehearse incident reporting before enforcement starts

PS26/2 enforcement begins March 2027. Your team rehearses the full detect-escalate-classify-report flow under realistic pressure. The worst time to learn your incident reporting process is during an actual incident.

Walk into the board meeting with proof

Every simulation run produces FCA/PRA-grade after-action documentation automatically. Board-ready, audit-ready, no manual formatting. A clear trail: what was tested, what broke, what's been fixed.

Build a living resilience programme

Run scenarios monthly, not annually. Compare results over time. Watch your response capability actually improve. The PRA's stress test is looking for proof that your organisation learns, and this gives you that evidence.

Built by people who've lived the problem.

Dr. Runli Guo

CEO & Founder

Founded AI Dionic in March 2024 after a 16-year cybersecurity career spanning payment services, gaming, insurance, and tech sectors. Former CISO at Gett and Valarian, Global Head of Digital Cybersecurity at Marsh McLennan. Named a Top 30 UK CISO in 2022 by CSO Online. PhD in Artificial Intelligence from the University of Surrey, specialising in Natural Language Processing.

Anthony Perridge

Chief Revenue Officer

Strong track record taking early-stage cybersecurity companies to international markets while building pipeline and developing partner programmes. Previously VP International (EMEA) at ThreatQuotient and EMEA Sales leadership roles at Sourcefire and Cisco Security.

Ray Stanton

Non-Executive Director Chair

Led the Global System Integrator Division at BT Global Services. Helped establish World Economic Forum cyber resilience programmes and sat on the Information Security Forum advisory board. Co-founded the White Hat foundation in 2004 and honoured with the Lifetime Achievement Award in 2025.

Dr. Ali Baghdadi

Board Advisor

Founded Aptec as a student, growing it to a $2 billion company before its acquisition by Ingram Micro, where he became CEO for the META region and SVP. Holds several pending AI and Cyber Security patents. PhD in Computer Science from Imperial College London.