Agent Workflow Failure Post-Mortem Communication Plan

Role: Principal AI Reliability Director managing high-stakes enterprise autonomous agent deployments.

Context

• Client Stakeholder: {{enterprise_client}}
• Incident Summary: {{incident_summary}}
• Impacted Agent Chain: {{affected_tool_chain}}
• Identified Root Cause: {{root_cause_analysis}}
• Technical Remediation: {{recovery_actions}}
• SLA & Business Impact: {{sla_impact}}

Task

Generate an incident communication plan and executive-level resolution email that details the root cause of an autonomous agent execution failure, outlines remediation steps, and restores client trust.

Method

1. Deconstruct {{incident_summary}} to isolate where tool selection, execution, or looping broke down.
2. Map the failure sequence across {{affected_tool_chain}} to provide a clear timeline of events.
3. Translate complex LLM or tool error modes into clear, business-appropriate language for {{enterprise_client}}.
4. Present {{root_cause_analysis}} highlighting guardrail adjustments and deterministic safeguards.
5. Detail the immediate {{recovery_actions}} taken to stabilize production agent chains.
6. Address {{sla_impact}} transparently, specifying monitoring enhancements and circuit-breaker policies.
7. Format the complete communication dispatch with clear executive summary and technical appendices.

Constraints

• MUST clearly distinguish between model non-determinism, infrastructure outages, and schema validation errors.
• MUST NOT shift blame to third-party model providers without articulating internal redundancy steps.
• Language must maintain high accountability and technical rigor.
• The email body must be under 450 words while technical appendices provide full context.

Output format

• Executive Dispatch Plan: Stakeholder mapping and delivery timing
• Client Incident Email: Executive Summary, Incident Timeline, Root Cause, Preventative Measures, SLA Next Steps
• Action Item Matrix: Remediation task list with target completion dates

Self-review

• Ensure the explanation of tool failure mechanisms is accessible to both CTOs and business leads.
• Confirm that every remediation step directly targets a factor identified in the root cause.
• Validate that all provided variables are contextualized correctly.