Video by The Linux Foundation via YouTube
Sponsored by Harness
AI can already write your Terraform. Point it at a requirement and it will generate the modules, the variables, the pipeline any code including infrastructure as code, at AI speed. That part’s solved.
What’s not solved is everything after we generate code: is this change compliant? What breaks if it’s applied? And when the real world drifts from what’s declared in code (ie a manual console change, an expired resource, a config someone "just fixed" by hand) who notices, and who fixes it? In this session, we’ll walk through the full arc of AI in Infrastructure Automation: from generating IaC, to governing what it’s allowed to do, to healing infrastructure when it drifts out of the state it’s supposed to be in.
We’ll discuss:
– How AI agents are moving beyond writing code into governing and maintaining it after it ships
– How the AI Blast Radius Agent analyzes a proposed change before it’s applied and brings true governance at the moment of decision, not after the fact
– How the Remediation Agent detects drift and automatically opens a fix PR giving the "heal" step, closing the loop without a human paging themselves at 2am
– Why an agent needs context, not just tools: how a Software Delivery knowledge graph connecting code, infra, deployments, and policy lets an agent judge whether a change is safe, not just execute it
– How governance is shifting from static, point-in-time approvals to continuous, policy-as-code verification
We’ll close with how organizations are approaching AI-native Infrastructure Automation with Harness IaCM and Harness AI to write, govern, and heal, on one platform, so infrastructure can move at AI speed without losing control of what actually gets deployed.