A renewal lands in the inbox. A 14% uplift, 12 pages, 30 days to sign. The finance lead does what finance leads have always done: asks around. A Slack message, a WhatsApp group of other CFOs, maybe a procurement consultant with a benchmark deck from last year.
Nobody knows. The only price your company has ever seen for this vendor is the one on your own contract.
Meanwhile the vendor's side of the table knows everything. They know what every customer pays, which discounts closed which deals, how hard each buyer pushed, and exactly how much uplift each account absorbed last year without complaint. B2B software pricing stayed opaque for 30 years because every buyer negotiated alone against a seller who never did.
That asymmetry is over. The system that pays your bills does not negotiate alone. It sits in a room with hundreds of other ledgers, and the same vendor appears in a great many of them, at a different price in each one. The information that decides a negotiation stopped being the vendor's private property the day spend became data.
Your accounting system already knows where you sit on that line. Vendor intelligence is the moment it starts telling you.
There are 3 questions your books can now answer that no salesperson will:
Question 1
Am I paying too much?
The same product, priced across an anonymised cohort of companies your size. Not a survey, not a consultant's estimate. The actual spread, from the actual ledgers, with your contract plotted on it.
Your percentile, not your gut
Question 2
What should I pay?
A renewal is a negotiation where 1 side has all the data. The counter-anchor is the cohort price, surfaced before you sign, not discovered a year later over drinks with another CFO.
The number to open with
Question 3
Should this vendor exist?
The newest question, and the sharpest. For a growing class of tools, the honest benchmark is no longer a rival vendor. It is what your own team builds with AI in a weekend.
Replace, not renegotiate
The third question deserves the numbers, because we ran it on ourselves.
Our marketing site lived in Framer. Our docs lived in Mintlify. Both had natural owners, both billed annually, both did their job, and neither survived contact with the question. We moved both surfaces into our own codebase, next to the product. The site got faster, the docs got better, and the vendors went away. The full story is in The Great Flattening.
We saved $15,000 to $20,000 a year on Framer and Mintlify alone.
The point is not those 2 tools. The point is that the deciding information, what this category costs elsewhere and what replacing it takes now, was sitting in the ledger and in the codebase the whole time. Nobody had connected them. Every vendor list has the same shape once you look: a few tools priced fairly, a few priced for whoever does not check, and a few that exist only because building was once harder than buying. Your books can already tell you which is which.
Connecting it is agent work, because the work is reading, and reading at scale is what agents do. A renewal used to be a document a busy person skimmed in the week it was due. Now it is a file an agent opens 90 days early:
- The contract: term, uplift clause, notice window, and the auto-renew date that used to pass silently.
- Your price against the cohort: the percentile you are actually on, not the "best and final" you were told.
- Usage against entitlement: the seats paid for versus the seats that logged in since spring.
- The uplift against history: what this vendor asked of comparable customers, and what they settled for.
- The replacement check: what this tool does, what building it takes now, and what the last company that switched saved in cost, speed, and quality.
Out the other side comes a brief, not a dashboard: here is the number to open with, here is the clause to strike, here is the walk-away alternative, drafted before the vendor's first email arrives. The negotiation still ends with a human signature. It just no longer starts with a human guess.
None of this requires anyone's secrets. The cohort is anonymised and aggregated, prices carry no names, and a benchmark only surfaces where the group behind it is large enough that no single company shows through. Every participant gives up nothing and learns what the market actually charges. The only party that loses information advantage is the one that was selling with it.
From here, the old ritual looks strange. Asking a WhatsApp group what a vendor should cost is the fax cover sheet of procurement: a workaround kept alive only because the real information had no way to reach you. It has a way now. It flows through the same system that books the invoice, pays it, and files the receipt, which is why the intelligence belongs there and nowhere else. Software that keeps your ledger has started reading it back to you.
The question is no longer whether your accounting system price-checks your vendors. It is whether it does so before your next renewal, or after you have signed it.
