Where Intervention Readiness matters most.
Systems are becoming more capable. Organisations are becoming more dependent on automation. Regulators increasingly require effective human oversight. The question is the same in every sector: can you stop it in time? Our initial commercial focus is large UK and European regulated financial institutions; other sectors are discussed case by case.
UK regulated finance
SS1/23 sets the expectation that model risk is managed and owned. SMCR names a senior individual accountable for it.
A named SMF can hold accountability for a model they cannot halt inside the harm window. The chart says control; the live event may say otherwise. AGDA™ assesses whether intervention remains possible.
EU high-risk AI
The AI Act requires human oversight of high-risk systems, capable of intervention and of overriding or halting the system.
Oversight that is informed after the decision has landed is documentation, not control. AGDA™ assesses whether oversight can alter the outcome before the consequence compounds.
Critical national infrastructure
Resilience regimes require that essential services withstand and recover from disruption inside defined tolerances.
Automated systems act faster than the escalation chain meant to govern them. When the chain runs slower than the harm, the tolerance is breached before anyone is authorised to act.
Public services
Automated decisions at population scale carry duties of fairness, transparency and an effective route to redress.
A redress path slower than the harm it is meant to contain is redress in name only. AGDA™ assesses whether intervention and correction land before the decision becomes irreversible for the person affected.
Is your situation in scope?
If your most consequential systems carry regulatory, financial, operational or reputational exposure, the question applies. Tell us the system and the decision it supports.