Who Owns Cloud Waste When AI Flags It?

Cloud cost dashboards have gotten faster, flagging anomalies within hours. Yet waste keeps rising anyway. The problem isn't visibility; it's who has the authority to act on what's flagged, said Sanjeev Mittal, chief product and technology officer at CloudKeeper.

"Someone who's looking at any optimization opportunity will end up raising a ticket, it goes into the backlog, and most likely it kind of dies there because nobody owns that cost optimization," Mittal said.

He traced the issue to a mismatch of personas: engineers can change the environment but don't own the cost, while finance owns the cost but can't touch the infrastructure.

On artificial intelligence autonomy, Mittal drew a line between two types of action. Commitment and rate optimization don't affect live infrastructure and suit AI decision-making well, he said. Usage optimization, which modifies the customer's environment, still needs policy-driven guardrails and human oversight, especially in production.

"The judgment moves into the tooling, but accountability does not move into the tooling," Mittal said of where FinOps is headed as cost intelligence gets built directly into engineering tools like IDEs. Rather than disappearing, he argued, FinOps is dispersing into engineering itself, much like DevOps reshaped software deployment.

In this video interview with ISMG, Mittal also discussed:

• Why cost per resource no longer works as a proxy for AI spend, and what should replace it;
• Where AI can safely act on cloud spend versus where it needs human-in-the-loop review;
• How MCP integrations are bringing cost intelligence directly into developer workflows.

Mittal has over two decades of global experience in cloud, SaaS platforms, enterprise software and product-led growth. He leads CloudKeeper's product and technology strategy, driving innovation in cloud optimization and scalable solutions that help enterprises maximize value from their cloud investments.

The interview was originally published on CIO.inc