Ambit Observatory

Ambit Observatory is a diagnostic and assurance layer built on top of Ambit Authority. It consumes the decision ledger produced by Authority and explains what happened, why it happened, and how governance policies behaved over time.

Authority decides. Observatory explains.

Availability

Observatory is offered to design partners during evaluation. General availability is intentionally limited while we harden the evidence surface with real workflows. If you need Authority, you have access to the decision ledger and evidence records regardless of Observatory availability.

Diagnostic Capabilities

Observatory is built for incident review, governance validation, and evidence-backed policy refinement — not generic performance dashboards.

Governed Traces

Trace any action back to the delegation, policy version, and decision evidence record that governed it. Investigations start from evidence, not logs.

Evidence Bundles

Time-bounded exports of decisions, delegations, and policy versions — structured for audit, incident review, and regulatory inquiry.

Governance Learning

Surface patterns in denials, escalations, and overrides that signal miscalibrated policy or excessive approval toil — so governance improves before the next incident.

Architecture

Observatory runs out-of-band, independent of any agent runtime or execution layer. It consumes Authority evidence records and governed traces; it does not gate actions. Because Observatory operates outside the critical path, it cannot affect governance decisions or action latency. The decision ledger is the source of truth.

Agent

intent

Authority

1 Evaluate policy + delegation 2 Render decision: ALLOW | DENY | ESCALATE 3 Commit receipt (append-only)

synchronous · before execution · eval time recorded

Tool / API

consequential action

Observatory

1 Ingest receipts / governed traces 2 Explain outcomes (RCA) + advisory signals 3 Recommend policy refinement (human-gated)

asynchronous · after decision

Diagnostic Outputs

Observatory produces tangible diagnostic artifacts derived from the Authority decision ledger. These outputs are advisory and explain outcomes; they do not replace Authority's evidence or governance decisions.

Governed Traces

A structured lifecycle of governed actions linked to Authority decisions, delegation state, and policy context.

Root-Cause Narratives

Explanations of unexpected outcomes grounded in what was authorised at decision time.

Decision Ledger

ground truth · append-only · tamper-evident

Advisory Signals

Patterns and recommendations that suggest where policy may be too strict or too permissive.

Evidence Bundles

Time-bounded exports for audit and incident review (decisions, delegations, policy versions, and outcomes).

Every output traces back to the Authority decision ledger. Nothing is inferred, estimated, or reconstructed from logs. The evidence surface is only as good as the decisions it can explain.

Observatory explains. Authority decides.

Authority deep-dive