Bank data tells us where to look. Statements tell us what to say.
Both matter, and they do different jobs. Understanding the difference is most of what separates a real analysis from a guess.
01
You connect your accounts
One read-only connection, about a minute.
We use Plaid, the same connection layer used by most business finance tools. Your banking credentials go to your bank, never to us.
The connection is read-only. It can see transactions. It cannot move money, and it never could.
You can disconnect at any time, and the data goes with it.
02
We find and size your vendors
Every recurring charge, sorted by what it actually costs you.
We identify recurring vendors by their bank descriptors and group them by category — card processing, payroll, shipping, software, telecom and more.
We separate what is negotiable from what is not. Your payroll wages and your payroll taxes are not savings, and we never count them as such.
You see annualized spend per vendor, ranked, so you know where the money actually is before anyone makes a phone call.
03
You send the statements that matter
This is where an estimate becomes a number.
Bank data has a hard limit: most processors deposit net of fees, and payroll providers pull wages, taxes, and fees in a single debit. The bank feed cannot separate them.
A statement can. It shows gross volume, transaction count, and every fee line — which is what makes an effective rate calculable and a markup visible.
We tell you exactly which document we need for each vendor and what it will let us prove. Upload it, and we read every line.
Then you choose how it gets fixed
You make the call
We give you the line items, the target number, who to ask for, and what they will say back. Plenty of owners prefer this, and it works.
We make the call
You authorize us to speak to the vendor on your behalf. We negotiate, you approve anything before it is sent, and we keep watching afterward so the rate does not drift back.