NEW REPORTAgent Down: The State of Resilience in Agentic AI Adoption is out now.Read the findings →

FRONTIER AI RISK

Research Board briefing Regulatory signals Simulations Community

FRONTIER AI RISK · FINANCIAL SERVICES

Is your resilience keeping pace with your AI?

As AI becomes embedded in critical financial services, new dependencies are emerging faster than organizations can test them.

Independent research, regulatory intelligence, and real-world simulations exploring what it takes to remain resilient as AI changes how financial services operate.

TRUSTED BY LEADING ORGANIZATIONS

INDUSTRY RECOGNITION

Different responsibilities. Shared exposure.

BOARDS & EXECUTIVES Can management demonstrate resilience under an AI-enabled operating model? Explore the governance decisions, emerging dependencies, and evidence boards need to oversee AI-related operational risk.

Agent Down: Board briefingFive priorities for agentic AI oversightRegulatory developments

Explore the briefing →

AI PRODUCT & ENGINEERING What happens when the AI you deploy becomes a critical dependency? Explore failure scenarios, operational safeguards, and the practical challenges of maintaining services when AI systems or providers become unavailable.

Designing credible fallback arrangementsTest your decisions in Agent DownAgent Down research findings

Explore the scenarios →Suggest a scenario →

RISK, RESILIENCE & SECURITY Are your recovery capabilities keeping pace with AI adoption? See how resilience leaders respond to AI-related disruption and where tested capability differs from expected readiness.

Agent Down research reportExperience the microsimulationUnderstanding Capability Intelligence

Explore the findings →

AGENT DOWN

FEATURED RESEARCH · OCTOBER 2026 The State of Resilience in Agentic AI Adoption

What happens when a critical AI provider becomes unavailable?

31 resilience leaders navigated a simulated failure of a critical agentic AI vendor. Their decisions reveal how AI adoption is reshaping operational dependencies, recovery assumptions, and confidence in maintaining critical services.

Read the report online → Download the PDF →

WHAT THE RESEARCH REVEALED

16%Were confident they could maintain minimum service levels following the AI vendor failure.

39%Turned to workforce-based recovery solutions.

77%Expected to breach impact tolerances sooner as AI increased operational volumes.

REGULATORY INTELLIGENCEAI is changing what supervisors expect organizations to demonstrate.

Across jurisdictions, regulators are sharpening expectations for operational resilience, third-party dependencies, oversight, and credible recovery arrangements.

Explore developments shaping how financial institutions govern, test, and prepare for AI-related disruption.

Explore regulatory developments →

TODAYOCCAPRAFCAOSFIMASESAsFSBJanFebMarAprMayJunJulAugSepOctNovDec

17 Apr · USOCC / FDICInteragency model risk guidance

30 Apr · AustraliaAPRACredible fallback for critical operations

6 Jul · UKFCAMills Review: assistance to delegation

13 Jul · CanadaOSFITest AI outage scenarios and manual fallbacks

31 Jul · EUESAsFrontier AI and ICT risk

7 Oct · SingaporeMASContingency for third-party AI failure

15 Oct · GlobalFSBSound practices for responsible AI adoption

CAPABILITY INTELLIGENCEResilience needs evidence.

Plans and assessments describe how organizations expect to respond. Testing reveals what people and teams can actually do.iluminr transforms decisions made during microsimulations into evidence of operational capability, helping organizations identify strengths, expose gaps, and prioritize improvement.Discover Capability Intelligence →

Simulations run, people exercised, objectives met, critical services exercised

EXPAND ↗

EXPAND ↗

EXPAND ↗

See Capability Intelligence on your own dataA 30-minute briefing on how your teams would score, and what the board would see.

Book a briefing →

EXPERIENCE THE SCENARIO The AI provider is down. What do you do next? Seven decisions. Competing priorities. A critical service at risk.Step into Agent Down and experience the decisions facing resilience leaders when a critical AI dependency fails.Approximately 15 minutes. Compare your decisions with those of the research cohort. Run Agent Down →

Explore the research

AVAILABLE NOW · OCTOBER 2026Agent DownThe State of Resilience in Agentic AI AdoptionExplore the decisions and findings from our first frontier AI risk simulation.Read the research →

IN DEVELOPMENTFrontier AdversaryOur next research initiative examines the evolving operational risks associated with AI-enabled adversaries.Get research updates →

UPCOMING DISCUSSIONAgent Down: findings, liveWith Marcus Vaughan (iluminr)Date TBC · OnlineGet the invitation →

New research and briefings, as they land.