To reliably prepare for the challenges of machine intelligence, the developer recommends implementing a comprehensive resilience assurance strategy, which encompasses four interrelated stages:
1. Recovery risk assessment. IT teams need to determine whether the current network configuration can withstand rapid cycles of vulnerability exploitation. The focus of the analysis shifts to the ability of clean recovery and verification of whether the environments for deploying backups are isolated from compromised production nodes.
2. Isolated recovery. Organizations must maintain reliable, immutable copies of critical data, fully separated from production accounts and management network planes. At the same time, continuous testing of targeted recovery time and point objectives (RTO and RPO) based on current attack scenarios is necessary, rather than just modeling technical hardware failures.
3. Prioritization by Minimum Viable Company principle. The strategy requires a precise definition of the systems without which the company cannot ensure viability: identification platforms, billing services, and basic cloud services. It is also important to consider new technological dependencies, such as vector-type databases and AI model repositories.
4. Automation and continuous testing. Companies need to set up automated threat scanning, precise identification of safe recovery points, and orchestration of processes in special isolated environments (“clean rooms”) even before a cyber incident occurs.