Harness has announced two new products designed to help organizations gain better insight into the costs and benefits of AI usage. The solutions address a growing problem within IT departments: AI is being deployed more frequently, but it is often unclear what those investments actually yield.
According to a Harness survey of engineering managers, 94 percent have not included cost metrics in how they evaluate software development, reports SD Times. As a result, there is a lack of visibility into whether AI tools contribute to higher productivity or, on the contrary, primarily generate additional costs.
The first new product, AI DLC Insights, focuses on software development. The solution tracks how many tokens developers consume via AI code assistants and links that data to concrete results, such as resolved bugs, pull requests, or new functionality. This allows organizations, for example, to determine how much AI capacity is needed to perform a specific task and whether that is more cost-effective than traditional development work.
Harness observes that developers are generating more and more code using AI, while some of it ultimately never makes it into production. The platform aims to reveal where AI investments actually lead to usable software and where money is wasted on unused code or inefficient processes.
The software supports multiple popular development tools, including GitHub Copilot, Cursor, Claude Code, and Windsurf. Additionally, costs and results are broken down by individual developers, teams, and business units. The platform also flags situations where token consumption increases without a corresponding rise in software production.
Insight into operational AI costs
In addition to development environments, Harness also aims to provide control over the operational costs of AI services. To that end, the company is introducing Cloud & AI Cost Management.
This solution consolidates AI spending from various vendors into a single overview. This allows organizations to track costs for services from providers such as OpenAI, Anthropic, AWS Bedrock, and Google Vertex AI. According to Harness, this overview is often missing, meaning that finance departments receive invoices but struggle to assess which applications add value.
The platform links expenses to specific AI agents, workflows, teams, and business units. This makes it clear which applications are working efficiently and which ones are actually incurring high costs without demonstrable results.
The solution also includes features for budget management and cost monitoring. Organizations can set limits for teams or projects and receive alerts when spending rises unexpectedly. This is intended to prevent a new AI application from quietly leading to a sharply rising bill.
According to Trevor Stuart, senior vice president at Harness, the biggest problem right now isn’t the level of AI spending itself, but the lack of insight into its return on investment. With the new products, the company aims to help organizations manage AI investments in the same way they manage cloud costs: based on measurable results rather than assumptions.