#20liveorchestrationAI-powered
Chief AI Coordinator
Orchestrate the agent panel and return the final combined report.
Responsibilities
- •Plan the pipeline
- •Run agents
- •Merge outputs
- •Resolve conflicts
- •Return the final answer
Inputs
- •Company profile
Outputs
- •Combined executive report
- •CIO decision
- •Overall confidence
Required tools
None — synthesizes other agents' outputs.
Permissions & access
Permissionagent:chief-ai-coordinator:run
Memory & context
Scopesshort_term, company, long_term
The company profile plus prior agents' outputs in the current pipeline; company-scoped long-term memory across runs.
Retry logic
Max attempts1
Strategynone
Backoff0ms
Confidence & human review
Overall confidence = min(CIO confidence, combined panel confidence).
Flag when any agent in the pipeline flagged review.
Cost optimization
Sum of the panel's agent costs; optimizes by reusing deterministic tools and gating the LLM behind a key.
All live agents (via the runtime)
Logging & monitoring
Structured AgentRun row: input, output, confidence, tool-run ids, duration, review flag.
success ratep95 durationconfidence distributionhuman-review rate
Future improvements
- •Parallel execution of independent agents
- •Token-budget-aware planning
- •Conflict-resolution voting
Pipeline position: Master orchestrator — runs the whole pipeline.
Run the full pipeline →API: POST /api/agents/chief-ai-coordinator