Kori sees what your team can’t.
The first AI FinOps analyst built to watch every line of your cloud bill, every day, and tell you exactly what to do about it.
m6i.4xlarge instances in eu-west-2, tagged dev-cluster. Want me to investigate?See it.
You can’t fix what you can’t see.
Know what moved, every morning.
Reports tell you which teams went up, which went down, and by how much. Before standup. Without anyone building a dashboard.
Daily spend — Tue 04 May
From a £40k anomaly to the exact resource, in two clicks.
Cost Explorer drills from total spend to the individual resource driving it. No spreadsheet exports. No tagging archaeology.
RDS — £42,180 this month
db.r6i.8xlarge — orders-primary
See why a machine costs what it costs.
Every resource breaks down by their subscosts. Storage, CPU, network, and IOPS, with infra metrics and tags alongside. So you optimize the expensive costs.
Fix it.
Seeing the waste isn’t fixing it.
Recommendations with the expertise baked in.
Rightsize this. Switch this from Windows to Linux. Combine these three idle servers. Wait two weeks before committing to that Savings Plan. Each one explained, costed, and ranked.
m6i.4xlarge → m6i.xlargeRight-size before you commit.
Most tools recommend Reserved Instances on top of oversized infrastructure, locking in the waste. Koritsu sequences it the right way: rightsize first, then commit. So you’re saving on what you actually need.
Every recommendation tracked to realised saving.
Suggested → in progress → live. Nothing quietly disappears. Your CFO sees the savings hit the bill, not just the slide deck.
Keep it gone.
Cloud environments change, waste grows back.
Anomalies hit Slack the morning they appear.
Not three weeks later in a finance review. The runaway batch job, the forgotten dev cluster, the misconfigured replica, flagged the day they show up.
m6i.4xlarge in eu-west-2, tagged dev-cluster.£412/day · platform-core · 08:14nightly-rebuild hasn’t terminated since Friday.£186/day · data-pipelines · 08:14us-east-1 serving zero traffic for 9 days.£94/day · checkout-svc · 08:14The report your CFO wants, in their inbox, every Monday at 9.
Scheduled reports to Slack, email, or Teams. Scheduled once, sent forever. Without you building it.
2. New anomaly in checkout-svc — investigation owned by jamie.s, ETA Wed.
3. Two Savings Plans expire in 23 days — commitment review scheduled.
Meet Kori.
Kori isn’t a chatbot pointed at your AWS bill. It’s a purpose-built FinOps analyst, trained on the patterns, traps, and trade-offs of real cloud environments, that lives inside your team’s day.
platform-core.m6i.4xlarge instances first appeared Tue 28 Apr 04:32.dev-cluster autoscaler change merged Mon 27 Apr.dev-cluster autoscaler started over-provisioning after Monday’s merge. The waste is £412/day and growing. Suggested fix: cap min instance count at 1 — saves the bulk of it without affecting the dev workflow. Want me to draft the PR?