Most cloud cost optimisation does not start with a grand FinOps programme.
It starts with a simpler question:
What are we still paying for that nobody is using?
We recently launched Cloud Cost Sentinel on TomorrowCentral to help with exactly that.
The first version is intentionally narrow: identify cloud resources that may be safe to review, shut down, resize, or remove. Think unused resources, stale infrastructure, and obvious cost leaks that often survive because nobody has time to investigate them manually.
The interesting part is not just the UI.
Cloud Cost Sentinel is also exposed through MCP, so an AI agent can call the same capability while investigating cloud spend. That means a person can use the console directly, or an agent can use the tool as part of a larger cloud review workflow.
This is where I think enterprise AI starts becoming practical:
- small bounded tools
- scoped access
- clear output
- human review before action
- agent-compatible interfaces
- measurable value
Not every AI use case needs to start with a giant autonomous system.
Sometimes the right first step is a narrow tool that helps a human or agent find waste faster.
Top comments (0)