Harness Inc. said today it’s injecting artificial intelligence into every stage of the software delivery process, looking to automate every step after the coding process, including testing, security, deployment and maintenance.

The new Harness AI suite builds on the company’s flagship platform, including the Harness Internal Developer Portal, which is designed to support continuous integration and continuous deployment or CI/CD methodologies, where companies push out frequent updates for their software and applications, often several times a day. But the AI suite is infused across the entire Harness platform and all of its modules.

In a blog post, Harness co-founder and Chief Executive Jyoti Bansal (pictured) said he wants the Harness AI platform to become the foundation for the next generation of software delivery, automating everything that comes after generating code to ensure new updates are pushed out faster, safer and more efficiently than ever before.

Bansal believes there’s an urgent need for automation across every step of the software delivery process. He points out that with AI-generated code now being used so extensively by developer teams, they spend little time actually writing code these days. As such, software delivery has become the biggest bottleneck in CI/CD workflows, with teams consumed by repetitive tasks such as managing pipelines, creating, running and maintaining tests and security scans, monitoring new deployments and analyzing costs.

Though some companies do use AI tools to help with these processes, it creates a fragmented experience that transforms CI/CD pipelines into complex patchworks that lead to new problems, with the result being all too frequent production failures.

“Harness AI is here to fix that,” Bansal said. “We believe that every stage after code generation — testing, security, deployments and maintenance — should be as fast, intelligent and automated as modern code generation. This is the next frontier of software delivery.”

Streamlined software delivery at every step

What the Harness AI platform does is it unifies intelligence access every step of the software delivery process, transforming those complex patchworks into a single, cohesive and adaptive system. Software teams interact with the platform through basic chat interface, and through this they can access an entire suite of intelligence AI agents, including DevOps, site reliability, release, app security and test agents that automate each of these important steps.

With Harness AI, almost every possible task involved in software delivery can be automated. Developers can use natural language to create a new CI/CD pipeline that automatically follows their company’s best practices. In addition, they can use it to automatically create, adapt and maintain tests, reduce costs by spotting inefficiencies and optimizing infrastructure resources, and validate each new deployment in production, or roll it back immediately and automatically if something goes wrong.

Bansal promised a uniquely seamless experience. For instance, if a user requests a new pipeline to deploy an application that’s in-line with their company’s standards and policies, Harness AI will understand that request and delegate the task to the DevOps Agent, which creates the necessary pipeline. If it needs help along the way, such as if it needs to scan the app it’s building, it will then ask the AppSec Agent to complete that task.

“You don’t see the agents, but you feel their impact in every workflow: faster, safer and smarter delivery,” Bansal said. “They’re the invisible force that powers every AI-driven capability in the platform.”

At its core, Harness AI orchestrates intelligent workflows that combine automated testing, deployment, infrastructure management, security, continuous verification, experimentation and cost optimization. A key element of this is the Harness Software Delivery Knowledge Graph, which records every change and aggregates insights across each of these steps to ensure a complete picture of everything that has happened. By ingesting data from the Harness IDP and other DevOps tools, the Software Delivery Knowledge Graph is able to map relationships among people, pipelines, services, incidents and infrastructure resources. That means each AI agent has full context around whatever task it’s assigned to complete.

“This secure, context-aware intelligence is what allows Harness AI to anticipate risks, automate workflows and optimize delivery with unmatched accuracy,” Bansal added.

According to Bansal, early adopters have managed to slash software testing cycle times by around 80% on average, reduce test maintenance times by almost 70%, while cutting the amount of time spent on debugging by about 50%.

Photo: Harness