We just released AI-powered remediation for the CxM platform. Before we built it, we spent months talking to engineering teams about why cloud optimization tickets never ship, and the pattern was consistent everywhere: excellent visibility into waste, zero bandwidth to fix it.
Every major FinOps platform can tell you where money is being wasted, down to the resource and dollar. Yet according to Boston Consulting Group’s 2025 report, Cloud Cover: Price Swings, Sovereignty Demands, and Wasted Resources, up to 30% of enterprise cloud spend remains wasted. The visibility exists, the waste persists. We built AI remediation to close that gap, and here's what we learned about why it exists in the first place.
What Remediation Actually Requires
Consider a straightforward optimization: right-sizing an autoscaling group running oversized instances. The dashboard flags it instantly, the savings are clear. But shipping the fix requires work that has nothing to do with identifying the opportunity.
Someone needs to trace ownership without reliable tagging, map dependencies to understand what might break, validate the change against deployment patterns, write the infrastructure code, document rollback procedures, get it reviewed and merged, then monitor post-deployment. That's three to five hours for one optimization, while the backlog has dozens more and the sprint is full of feature work and incidents.
When we asked engineers what prevented them from shipping optimizations, the answer wasn't about caring about costs or knowing what to fix. They simply didn't have five extra hours for archaeological research.
Why the Same Waste Appears Every Quarter
Most organizations operate reactively. Monthly reports surface problems that have been accumulating, tickets get filed, sprints fill, and those tickets sit. Next quarter brings the same report, same recommendations, same waste still burning money.
This happens because optimization work arrives as research projects competing with feature delivery. Engineers see tickets that will take hours of investigation and deprioritize them, which is a reasonable decision given their constraints. But the waste keeps accumulating while the tickets age out.
Building CxM and talking to our customers, we realized the problem was that remediation fundamentally doesn't fit into how engineering teams work. You can't sprint-plan five hours of archaeology per optimization when you have dozens of optimizations and a roadmap to ship.
What Proactive Means in Practice
Proactive remediation inverts the model. Instead of discovering waste and then scrambling to fix it, you prevent waste from accumulating in the first place by identifying optimization opportunities continuously rather than monthly, validating and preparing fixes automatically rather than manually, and delivering remediation as executable work that fits into existing workflows rather than research projects competing with feature delivery.
An optimization that requires five hours of context-gathering will sit in a backlog, while one that requires five minutes of code review will ship. Making remediation manageable within the bandwidth engineering teams actually have changes what gets done.
When remediation becomes proactive, waste gets eliminated before it accumulates rather than being discovered after the damage is done. Optimization stops competing with feature work because the research is already complete, and progress becomes measurable in prevented costs rather than theoretical savings in retrospectives.
Where AI Actually Helps
AI solves the remediation gap by automating the work that makes remediation so time-intensive. Our AI traces ownership across thousands of resources without manual tagging by inferring relationships from deployment patterns and infrastructure dependencies. It validates changes safely by understanding system context and predicting impact, generates correct code modifications by learning infrastructure-as-code patterns and organizational conventions, and improves precision over time by tracking which recommendations teams accept and which get rejected.
The AI handles the archaeological research and validation work while engineers focus on the decision that matters: does this change make sense for our system? That becomes a five-minute review instead of a five-hour investigation.
What Changes
When remediation becomes proactive and automated, the economics shift entirely. Optimization stops being a quarterly fire drill and becomes continuous, the backlog of unaddressed waste stops growing, and the gap between identified savings and realized savings closes.
Engineers stop seeing optimization as a tax on their time, FinOps teams stop generating recommendations that die in backlogs, and managers stop watching the same waste appear in consecutive reports.
The waste problem has persisted because fixing it requires more bandwidth than engineering teams have. We've integrated AI-powered remediation into the CxM Platform to change that equation.
Read more about the release here: AI-Powered Remediation Now Available
