About
Cost clarity for the AI era
Cloud bills used to be the hardest line item in engineering to explain. Then AI APIs made it harder. IRIS exists so engineering and finance teams can see the full picture — AWS and AI providers — in one place, and ask plain-English questions of the data instead of waiting on a query.
Why we built it
Modern cloud bills aren't just AWS anymore. A typical Series A or B startup runs infrastructure on AWS and ships features powered by OpenAI and Anthropic. Each provider has its own console, its own export format, its own delay. Finance ends up reconciling three CSVs at month-end. Engineers find out about a spike a week after it happened.
The tools that exist today were built for a world where cloud cost meant infrastructure cost. IRIS is built for the world where it means infrastructure plus AI — and where the most valuable thing you can do with that data is ask questions of it in real time.
What we believe
One pane of glass
AWS and AI provider spend belong in the same view. Splitting them across three dashboards is how surprises happen.
Read-only by default
A cost tool should never need write access to your infrastructure. We don't take AWS keys; we use a read-only IAM role you control.
Plain-English over SQL
Most cost questions are simple. They shouldn't require a data engineer. Our chat is essentially a FinOps engineer on call.
Stay in the EU
Customer data lives in AWS eu-west-1. Nothing leaves the EU through IRIS itself.
The team
Adam was Head of Engineering at a UK challenger bank, and most recently a Senior Lead Software Engineer at JP Morgan. At both, the same pattern played out: engineers spun up infrastructure without knowing what it would cost, and finance discovered the bill weeks later. AI APIs only sharpened the problem — teams burning through OpenAI and Anthropic tokens with no way to tell who was responsible, on which model, or for which feature. He founded IRIS in 2026 to close that gap.
LinkedInThe company
Want to talk?
We respond to every email personally.