WayFinder by Dojo Labs
AI Incident Response. Inside Your AWS. Deployed in 2 Weeks.
Alert ingestion pipeline
Ingests from New Relic, CloudWatch Alarms, and custom webhooks. Classifies signal vs. noise in seconds, eliminating the ~70% of alerts that do not require human attention.
Root-cause reasoning
Amazon Bedrock powers deep analysis across logs, metrics, and dependency maps. Identifies probable root cause, contributing factors, and blast radius in under 30 seconds.
Risk-tiered decision bands
Each incident is assigned to Band 1 (auto-fix), Band 2 (1-click approval), or Band 3 (full context handoff) based on confidence score and blast radius. No playbooks required.
Remediation and audit trail
Band 1 actions execute immediately via Lambda and Step Functions with defined rollback at every step. Band 2 delivers a Slack approval card. Band 3 delivers full diagnostic context to your team in under 30 seconds.
Slack approval workflow, phase 2
Band 2 one-click approval cards with confidence score display, blast radius summary, and audit-ready action trace. In pilot with a US regional bank this week.
Band classification engine GA
Risk-tiered routing logic assigns every incident to Band 1, 2, or 3 based on Bedrock confidence scores and configurable blast radius thresholds. Replaces the previous manual playbook approach.
CloudWatch + New Relic ingestion complete
Full coverage of CloudWatch Alarms, New Relic webhook, and custom sources. Output normalized to unified incident schema ready for the analysis stage.