We built an AI cloud engineer that prevents and fixes cloud waste autonomously.
The problem we kept seeing: FinOps tools generate hundreds of "optimization recommendations" that sit in dashboards forever. Industry implementation rate is ~5%. The tools work fine—humans just can't keep up.
So we asked: what if the tool actually fixed things?
Yasu connects to your AWS/GCP/Azure, identifies waste (idle resources, oversized instances, misconfigurations), and either auto-remediates or creates PRs for your review. We also integrate into CI/CD to catch costly mistakes before deployment—where fixes are 10x cheaper.
Tech stack: Multi-agent AI architecture, integrates with GitHub/GitLab for shift-left, lives in Slack/Teams for queries.
Early results: 30-35% cost reduction, 95% implementation rate.
We're a small team out of Utrecht (Netherlands). Would love feedback from anyone who's dealt with cloud cost pain—what's worked, what hasn't? / FinOps practice tips or IRL optimization tips!