Live Panel with
summary
AI agents automate incident ops by ingesting onboarding troubleshooting docs and extracted query semantics to rebuild effective DB queries, and by checking repo-checked memories/skills plus an agent-legible dependency graph to limit context growth. Teams deploy orchestrators that route specialized agents (hardware, network, software) and then distill recurring fixes into deterministic code for cost efficiency. Examples: noncritical bugs drop from 7 days to 7 hours, data-center technicians use agent-guided wiring checks, networks remain the common bottleneck.
AI-written summary · about
chapters (17)
- 0:00Agentic operations at scale
- 1:24Context engineering for AI agents
- 3:10Skills in monorepos
- 5:15Incident management agents
- 6:58Measuring AI incident response
- 9:27Deterministic vs AI knowledge
- 10:50Failure modes in AI infra
- 12:35Hardware vs software cycles
- 15:05Observability cost efficiency
- 17:35AI agents as risk mitigation
- 18:41Safety frameworks in AI
- 20:46Customer guarantees with AI
- 22:10Preventing cascading failures
- 22:53AI agents for reliability
- 24:38Networking in AI infra
- 27:31Innovation vs reliability
- 29:55Closing remarks
related talks
15mIntelligent Incident Management in DevOps
16mAI Agents for Incident Investigation: The Good, The Bad, and The Ugly
35mAgentic Incident Response
17mFighting AI with AI
46mLeveraging AI and Automation in DevOps
40mWhen Incidents Fix Themselves: AI SRE in action
32mWhat Agents Do When You’re Not Looking
24mAI SRE and Agent Handoffs: incident.io's Collaboration Model