Senior Director - Product
Ronak has over 11 years of experience in building AI/ML and data products and scaling engineering teams at various startups. He was part of the early teams at Cogoport, Jugnoo, and Peak AI.
2
2For a decade, cloud cost management has meant one thing: open a dashboard. Log in, filter, pivot, export. The problem was never the dashboard; it's that almost no one opens it until the bill is already too high.
Meanwhile, the way we work has changed. Engineers, platform teams, and FinOps practitioners now spend their day inside AI assistants - Claude, ChatGPT, Cursor, and Claude Code. That's where the questions get asked. It's not where the cost answers live.
Today, we close that gap. LensGPT - the AI cost analyst already used across CloudKeeper’s 400+ customers is now available as an MCP server.
Connect it to Claude Desktop, Claude Code, Cursor, ChatGPT, or Kiro, and ask about your cloud spend the way you’d ask a teammate:
“What did we spend on EC2 last month, and how does it compare to the month before?”
You get an answer grounded in your real billing data, not a guess - without leaving the tool you’re already in. RI coverage, Savings Plan utilization, cost breakdowns, the right dashboard - all conversational.
The hard part of putting an LLM near your cost data isn’t access. It’s trust. LensGPT is built for it: every number maps to a real query against your data - if the data isn’t there, it says so, it never estimates. Your data stays scoped to your account, and it’s strictly read-only. It analyzes spend; it can’t touch your environment.
If you’re already a CloudKeeper customer, you can turn this on today here; nothing new to instrument: the intelligence you already have, now where you work.
This is the start of a larger shift. Cost intelligence shouldn’t be a place you go. It should be ambient - in the conversation, then in the workflow, and eventually acting on what it finds, not just reporting it. The dashboard isn’t going away. But it no longer has to be where FinOps begins.
