Harness Introduces the Infrastructure Efficiency Loop, Turning Cloud Waste into Reviewed Infrastructure as Code Changes

Harness's Cost Management and Software Delivery Agents open pull requests for cloud savings, and Infrastructure Blast Radius scores each change's risk before engineers approve it

SAN FRANCISCO, Nov. 12, 2026 /PRNewswire/ -- Harness, the Autonomous SDLC Platform™ company, today announced general availability of the Infrastructure Remediation Agent and Infrastructure Blast Radius Agent. Paired with the Harness Cost Management Agent, the Remediation Agent creates the Infrastructure Efficiency Loop, a continuous workflow that takes cloud waste from finding to fix without a ticket in between, while Blast Radius scores the risk of each change before it is approved.

AI is accelerating software development, and every new application needs infrastructure underneath it. Platform teams are provisioning faster while under pressure to hold cloud spend down, and the Harness 2026 State of AI in FinOps report puts waste at 26% of cloud and AI spend.

Savings opportunities are easy to find and slow to act on. Each one becomes a ticket, an engineer translates it into infrastructure code, and the fix competes with everything else on the backlog. Tools that change cloud resources directly skip the queue, but they can leave live infrastructure out of sync with the code meant to manage it.

The Infrastructure Efficiency Loop

The Infrastructure Efficiency Loop makes optimization a continuous part of how engineering works. It runs in four stages:

• Identify: The Cost Management Agent surfaces cost optimization and drift findings.
• Remediate in code: The Infrastructure Remediation Agent turns a finding into an IaC pull request.
• Validate and apply: A pipeline checks the plan and Infrastructure Blast Radius scores its risk. Engineers review and merge, and existing pipeline controls govern how the change is applied.
• Track and repeat: The Cost Management Agent tracks which recommendations have been applied and keeps surfacing new opportunities as infrastructure and costs change.

"I'm most excited about the potential to turn these cost management insights directly into action. Instead of stopping at a recommendation we send out via email or Jira, having the platform create a pull request with the proposed remediation for our application teams changes the experience entirely. In our environment, these pull requests trigger a plan in Infrastructure Deployments, which will help give the context engineers need to confidently review and implement the change," said Dustin Mason, Principal Cloud Engineer at Citizens Bank.

New Capabilities For Platform Teams

The Cost Management Agent now works inside Harness Infrastructure Deployments, bringing the costs an organization pays, including negotiated discounts and enterprise agreements, into infrastructure workspaces. Cost optimization and drift findings appear in workspace Insights, next to the infrastructure engineers already manage. The agent infers applied status automatically for most implemented recommendations, and users can mark others as applied.

From any finding, the Infrastructure Remediation Agent generates an IaC change and opens a pull request. A pipeline can run plan-based checks and write the results back as pull request comments, and engineers keep control of the merge.

To safeguard these changes, the new Infrastructure Blast Radius Agent analyzes Terraform and OpenTofu changes and provides approvers with a risk score from 1 to 10, a plain-language explanation of the risk, and a dependency graph of the resources behind the change. Teams can use the score and risk level to require additional approvals for higher-risk changes, so the review effort goes where the risk is.

Better Together On The Autonomous SDLC Platform

"We're launching the Infrastructure Efficiency Loop so platform teams can finally act on cloud savings without trading away control," said Patrick Brogan, Field CTO at Harness. "Our customers already know where their waste is. What slows them down is the handoff between a finding and a fix, and that is the work we are taking off their plates. Engineers stay in charge of every change, and the savings move at the pace their teams ship."

The Harness Cost Management and Software Delivery Agents draw on the SDLC Knowledge Graph, the shared context across an organization's services, pipelines, deployments, and policies. That context lets the Cost Management Agent point to the right resources and lets the Infrastructure Remediation Agent write changes that fit the workspace they land in. Every action runs under the same identity, policy, and approval controls as the rest of the delivery process, so teams can move faster without giving up control of how changes are made.

Availability

The Infrastructure Remediation Agent, cost context in Harness Infrastructure Deployments workspaces, applied-status tracking in the Cost Management Agent, and Infrastructure Blast Radius are generally available today. See Harness Infrastructure Deployments in action by requesting a demo.

About Harness
Harness is the Autonomous SDLC Platform company, enabling engineering teams and agents writing code to deliver software quickly, reliably, and securely. Harness's platform includes the Software Delivery, Security Testing, Runtime Protection, and Cost Management Agents. The Agent Harness ties them together, sharing context from the Software Delivery Knowledge Graph and governing every action through identity, permissions, policy, and verification. Companies like United Airlines, Morningstar, and Choice Hotels use Harness to accelerate releases by up to 75%, cut cloud costs by up to 60%, and improve efficiency by up to 10x. Based in San Francisco, Harness is backed by Goldman Sachs, Menlo Ventures, IVP, Unusual Ventures, and Citi Ventures.

SOURCE Harness