StackGen puts one governance plane under every production agent
Coding agents made change volume explode. Operations absorbed the blast without a comparable upgrade — and StackGen's answer is to stop bolting agents onto four silos and put one operating system underneath all of them.
StackGen launched its Autonomous Operations Factory on Tuesday, a control plane that runs infrastructure, DevOps, observability and reliability agents against a single shared model of what is deployed, what changed, what broke and what fixed it. The foundation is Aiden OS, which pairs the Aiden World Model — a continuously updated knowledge layer spanning the operations lifecycle — with an agent harness that enforces policy, logs every action and routes approvals. Four first-party agents ship on top: Aiden for InfraOps, DevOps, SRE and Observability. Because they share the harness, what one agent learns is available to the others, an SRE investigation opens with deployment history already loaded, and a release is gated against live infrastructure drift before it ships. Customers can bring their own agents onto the factory and inherit the same guardrails and audit trail instead of rebuilding governance per agent.
The pitch rests on a real diagnosis. AI compressed the software development inner loop — more changes, more frequent deploys, bigger blast radius — while the outer loop stayed split across four functions with separate tooling and separate context. Point agents bolted onto each silo make each silo faster in isolation; they do not connect them. StackGen's CEO Sachin Aggarwal frames it as a factory needing an operating system: "one world model, one harness, one governance model." That is roughly what we saw a credit bureau build for itself in Why a credit bureau built the cleanest enterprise agent stack in 2026 and what Experian productized two weeks ago in Experian ships an Agent OS with ServiceNow as its first partner — the agent-governance layer is becoming a category, not a feature.
The detail worth noticing is the Autonomy Index, benchmarks delivered inside the product that score how autonomous each stage of the factory is, by team and by application type. That is the honest part of the offer, and the part that will matter commercially: enterprises that cannot say how much authority their agents actually hold cannot defend the rollout to an auditor or a board. It is also the shape this market has converged on. The infrastructure an agent inherits is the product now — the model inside it is interchangeable. Standalone point agents are a feature; the harness, the world model and the audit trail are the business.
Autonomous Operations Factory is in preview for AWS, Azure, Google Cloud and Oracle Cloud users, with all four agents available and the reliability agent also shipping in a free community edition. OneTrust is among the customers already using parts of it for observability and incident response. Access is by request, and StackGen has given no date for general availability.
What to watch: whether the Autonomy Index gets adopted as a buying criterion — if procurement starts asking for an autonomy score, StackGen set the units of the next enterprise budget line.
Would you let an agent gate a production deploy on drift it detected itself? Tell us in the comments.
