We do one thing: find money in corporate Snowflake accounts, and prove where every dollar of it came from.
Abr is Persian for cloud. A lens is what you reach for when the detail matters more than the impression — your cloud bill has both, and only one of them shows up in the dashboard.
Most cost reviews arrive as a slide with a big number on it and no way to check the arithmetic. Ours arrives as a register: one row per finding, each with the annual dollar value, the assumption it rests on, the change required, and who has to approve it. If you disagree with a number, there is something specific to disagree with.
That format exists because the alternative fails in a predictable way. A data leader can't take an unsourced figure to a CFO, and a platform team will reject any finding that misreads how their estate actually works. So we price from billed dollars rather than credits times an assumed rate, we reconcile storage against account totals before analysing a single table, and we write down what we checked and found healthy as carefully as what we found broken.
We also publish our retractions. On the engagement in the findings register we called storage clean, reversed it, and reversed it again before the reconciliation closed on a finding worth $5,800 a year. A review that never corrects itself hasn't been looking hard enough.
The audit is deliberately narrow and deliberately cheap to say yes to: three weeks, read-only, fixed fee, no access to your data. Most of what we find is configuration your own team can apply without us.