Let’s face it, AI is getting implemented almost anywhere. For the brave, there are now AI-powered platforms for infra, operations and security. Here’s a bunch of them and their marketing claims:
Incident Response is Getting an AI Copilot
Rootly AI claims up to 91% faster incident resolution. How? Auto-detecting root causes, suggesting remediation steps, and even drafting post-mortems while you’re still fighting the fire.
incident.io’s AI SRE does initial incident analysis in 1-2 minutes - what typically takes humans 15 minutes. It autonomously investigates, correlates recent deployments with errors, and surfaces relevant context.
Kubernetes Diagnostics Going Autonomous
K8sGPT (now a CNCF Sandbox project) analyzes your cluster, explains what’s wrong in plain English, and suggests fixes. Think of it as having an SRE who never sleeps and has read every K8s issue on GitHub.
Kagent takes it further - a CNCF project building an AI agent framework specifically for Kubernetes operations. Agents that can actually execute remediation, not just suggest it.
The Most Interesting Development: AI Operations Platform That Ships Code
Plural has AI agents that don’t just detect problems - they fix them. Autonomous code changes that create PRs for infrastructure modifications. Detect an anomaly? The AI drafts the fix and opens a pull request for review.
Harness AIDA helps debug failed builds, explains security vulnerabilities, and suggests remediations. Their ML-powered rollbacks automatically detect deployment issues and revert before users notice.
Have you tested any of these? Is it something you intend doing?
Originally posted on LinkedIn.