What it is
Signals are the atomic units of business process execution in AI systems. Each signal represents one verifiable step: a user request enters the system, triggers a policy check, invokes an AI model, calls a tool, awaits approval. Like droplets in a cascade, each signal can trigger subsequent signals — forming chains of cause and effect that mirror your business logic.
Why it matters
Business processes require consistency. When an employee processes a credit approval differently each time, chaos follows. AI agents face the same challenge — without signals capturing each step, there's no way to ensure consistent execution of business procedures. Signals turn 'the AI did something' into 'here's exactly what happened, in what order, for what reason.' They transform unpredictable agent behavior into auditable business processes.
Where it lives in AIOP
Every signal flows into the Audit Stream as it happens. The Correlate primitive joins signals by request ID into complete, replayable timelines. Higher-level artifacts — evidence packs, replay sessions, audit reports — are all built from these signal chains. No signal means no proof; every business-critical action must emit one.
Transform AI from unpredictable black box to auditable business process.
- Reduce operational errors by enforcing consistent execution paths.
- Enable process optimization through complete visibility into AI decision chains.
- Prove compliance with documented signal trails.
Finally see what AI agents actually do.
Have the audit trail they need.
Debug issues by replaying signal chains.
Optimize processes with complete execution data.