The tools as a network,
not a collection.

Every agent run in this system writes a timestamped record to the audit log: which agent ran, what it produced, whether it was blocked or flagged, and whether a human reviewed it. This panel is a live view of that stream — what the observability layer looks like in practice.

In an AI-native workflow, a PM can't personally review every agent decision the way they reviewed every document they wrote. The decision log restores that visibility. When an agent flags a WARN, a PM knows to look. When a HITL gate pauses the pipeline, the log records who was waiting and why. When a run is BLOCKED by policy, there's a record — not a mystery. This is what makes it safe to delegate synthesis to agents: you always know what they did and why they stopped.

• Retrospective review: "Why did the intake chain recommend REJECT for this feature?" — the log shows every step and every input.
• Stakeholder accountability: Named HITL sign-offs are recorded with timestamps, not just implied by an email chain.
• Quality signals: WARN entries flag AI outputs that need PM attention before flowing downstream — catching low-confidence claims before they reach a sponsor.
• Audit readiness: Every approved business case, signed intake output, and gate decision is timestamped and attributable to a named person.
• Cost visibility: Token spend per agent run is tracked — AI operational costs become visible, budget-able, and attributable to the team or feature that incurred them.