Engineering-grade FinOps
The FinOps consultant for when "why is AWS so expensive" reaches the board
Most FinOps work trims the purchasing layer and stops. Koritsu goes deeper, to the design decisions driving the spend, and gives you a real answer to take upstairs. Every saving verified against your actual bill.
What a FinOps consultant actually does
A FinOps consultant helps engineering and finance teams understand where cloud spend goes, who owns it, and where there is room to optimise. In most cases that work covers reservations, rightsizing, tagging and reporting. It realises real savings, and it is worth doing well.
The larger opportunities tend to sit one layer deeper, in how the system was built. A service running more often than it needs to. A cache that was never added. A non-production environment running close to production cost. These are design decisions, not purchasing decisions, and most engagements rarely get near them.
The difference in scale can be significant. Rightsizing trims a line item. A change at the design layer can remove most of the cost behind it. There are cases where a single finding took spend down substantially, because the resource was doing work it never needed to do. Koritsu is built for leaders who suspect the bigger savings are in how things are running, not in what they bought.
Where most FinOps stops
The purchasing layer
Buying decisions on top of the spend you already have.
Where Koritsu works
The architecture layer
The design decisions that create the spend in the first place.
Engineering-grade, not dashboard-deep
The difference is resolution. We start with billing data and infrastructure metrics, and move into the code or architecture only when the data says it is worth it.
What your bill doesn't show you
Cloud providers report cost by service, not by individual resource. That detail comes from exported billing data, often left switched off. We work from it, so a finding names the actual machine, not the category.
Cause, not symptom
Findings point to the code path, service behaviour or design decision driving spend. Missing caches. Inefficient timeouts. Idle capacity behind a steady bill.
Correct sequencing
Fix the pattern before you commit to a Reserved Instance or Savings Plan. Rightsize first, so you don't lock in spend you could have removed.
Verified against billing
Savings are measured against actual billing data, not projected on a slide. If it does not show up on the bill, it does not count.
See why a machine costs what it costs
Every resource breaks down into its sub-costs, storage, compute, network and IOPS, with infrastructure metrics and tags alongside. So when a finding shows storage is 62% of a database and the processor isn't the problem, the optimisation goes where the spend actually is.
That resolution is what lets a finding name a cause instead of a category, and what makes the estimated saving something you can stand behind before any change ships.
Visibility from day one, not month two
Most engagements start with a setup tax. The consultant learns your cloud's native tooling, builds a dashboard, or has you buy one and wait for it to populate. That is often one to two months before anyone sees the environment clearly. The clock on savings only starts after that.
Koritsu arrives with the platform already built. We are the consultants who bring the dashboard, not the ones who build it on your time. Clear visibility from day one, so the analysis starts immediately.
Time to first visibility
Weeks 1–8 reclaimed. The platform is already built, so the analysis — and the savings clock — start immediately.
A consultant, a platform, and an AI agent
Many teams already have a FinOps tool. Often a dashboard that rarely gets acted on once it is set up. Worth asking what it actually fixed last quarter. And a dashboard, bought or self-built, carries its own running cost. Maintenance, testing and hosting that rarely gets counted against the savings it is meant to produce.
| A FinOps dashboard | Koritsu | |
|---|---|---|
| Surfaces spend data | Yes | Yes, at resource level |
| Time to first visibility | 1 to 2 months to build or buy | Day one, platform already built |
| Names the architectural cause | Rarely | Yes |
| Sequences the fix correctly | No | Yes |
| Verifies savings against billing | No | Yes |
| Running cost of the tool itself | Maintenance, testing, hosting — often uncounted | Included, nothing to maintain |
| You pay only when savings are real | No, paid regardless | Success fee on verified savings |
Koritsu pairs the analysis with Kori, our AI FinOps agent, so the monitoring continues after the engagement. The next anomaly surfaces in days, not at the next quarterly review.
How an engagement works
No upfront fee. We work on a success fee only. The Savings Opportunity report is the core deliverable, and the fee follows the savings, not the report.
Scoped, read-only access
An explicit allowlist policy your security team can review and scope. Legacy and mid-migration accounts excluded from the start.
Savings Opportunity report
We analyse the environment and deliver specific, sequenced findings. Each one names the cause, estimates the saving and orders the work. Not a dashboard — a report you can act on.
You realise the savings
You decide what is worth acting on. We support the changes through to the bill. Savings are tracked against actual billing data, not projected on a slide.
Success fee on verified savings
The fee is two months of realised savings, measured against actual billing. If the bill does not move, there is nothing to pay.
Ongoing support, if useful
Convert into a monthly subscription for continued monitoring and reactive or embedded support. Or take the findings and run.
From platform-only to a fractional FinOps team
Different teams need different depth of support. The platform stands on its own; the people layer scales up from there.
Monitor
Platform only. Continuous visibility and anomaly alerting through Kori.
Advisor
Platform plus reactive FinOps support when you need a second opinion.
Embedded
A proactive, fractional FinOps team. Outsourced FinOps without a permanent hire.
Premium
Fully bundled. Platform, agent and team working as one function.
Common questions
A tool surfaces the data. The question is what was fixed as a result. A dashboard also carries its own maintenance, testing and hosting cost, rarely counted against the savings. Koritsu interprets the data, names the cause and sequences the fix, then verifies the saving against billing.
From the first week. A traditional engagement spends one to two months building or buying a dashboard before visibility begins. We arrive with the platform already built, so the analysis starts on day one.
No upfront fee. We work on a success fee only. The fee is two months of verified realised savings, measured against billing. The Savings Opportunity report is the core deliverable. Ongoing support converts into a tiered monthly subscription.
Read-only, scoped by an explicit allowlist your security team can review. We exclude legacy and mid-migration accounts from analysis and from savings figures.
AWS, GCP and Azure, and any provider that exposes FOCUS billing data.
Find out what your cloud is really costing you.
A 30-minute call. We'll look at your setup, share what we typically find in environments like yours, and tell you honestly whether we can help. No pitch deck. Whether you want a hands-on engagement or just want Kori watching your account, it starts here.
Book a discovery call