AI & Tuning

Fine-tune every threshold, model and AI autonomy level from one place · changes are versioned, back-tested and reversible

Security CopilotL0 suggestL1 approveL2 guardrailsL3 autonomous

Triage agentL0 suggestL1 approveL2 guardrailsL3 autonomous

Fix agentL0 suggestL1 approveL2 guardrailsL3 autonomous

Correlation agentL0 suggestL1 approveL2 guardrailsL3 autonomous

Document intake AIL0 suggestL1 approveL2 guardrailsL3 autonomous

L0 suggests, L1 acts after a person approves, L2 acts within guardrails and reports afterwards. L3 is blocked until the AI committee approves it.Governance →

Lifecycle

Security programme · step 1 of 7