TallyUp 3 min read

The knowledge engine · Part 2 of 7 · Method

What a contract actually knows

A signed contract is the most knowledge-dense document a business produces — and the first thing most systems do is shred it. What stays possible when the knowledge stays attached.

  • contracts
  • operations

Open the last contract your company signed. Not the summary — the document.

It knows who the parties are, and in what capacity. It knows what was promised, in both directions: the service to be delivered, the money to be paid, the dates by which each must happen. It knows the price, the currency, the unit the price is quoted in. It knows what happens if usage crosses a threshold, what happens if delivery slips, what the renewal looks like and when the window to decline it closes. It knows who carries which risk, and it knows — in the signature block — exactly who agreed to all of it.

A signed contract is the most knowledge-dense artifact a business produces. Pound for pound, nothing else comes close.

Now watch what happens to it.


The contract gets countersigned, and the shredding begins — not of the paper, of the knowledge.

The amount goes to the billing system. The dates go to somebody’s calendar. The renewal clause goes nowhere, until it matters. The payment schedule becomes a row per invoice, each row carrying a number and a date and none of the reasoning. The threshold terms go into the head of the account manager who negotiated them. The document itself goes into a folder, where it becomes the thing someone has to go find eight months later, when a payment arrives short and nobody remembers the credit it nets against.

Every downstream system receives its shred and does its job with it, correctly. The invoice is right. The journal entry is right. The forecast cell is right, for what it was given.

What none of them received is the meaning — the connective tissue that made each number a consequence of an agreement rather than a free-floating fact.

The invoice is not the knowledge. The invoice is the contract, with everything that made it intelligible stripped off.

So the knowledge gets rebuilt, by hand, on demand. Why is this payment short? Someone finds the folder. Will this revenue repeat next year? Someone reads the renewal clause. What are we actually exposed to if this customer churns? Someone builds the workbook. The answers exist — they were all in the document on signing day. The work is re-deriving what the business already knew once.


Here is the move that changes the arithmetic: keep the contract’s knowledge attached.

Not “store the PDF” — folders already do that. Keep the facts the contract asserts as facts in the record: the parties, the obligations in each direction, the schedule, the price and its currency and its unit, the renewal terms, the thresholds — each one connected to the others, and each one connected forward to everything that later happens because of it.

Then the invoice is no longer a shred. It is the contract, continuing to speak. The payment that settles it carries the chain back to the term that produced it. The short payment is not a mystery; it nets against a credit the record can point to. The renewal is not a calendar surprise; it is a fact with a date and a consequence, visible the whole time. The forecast does not assume the revenue repeats; it reads the terms under which it would.

None of this asks anyone to be more diligent. The knowledge was always there. What changes is whether the system keeps it or shreds it.


One more thing worth being precise about, because it is the difference between a record and a pile of custom fields.

The facts a contract asserts — parties, obligations, prices, currencies, dates — are not exotic. They are so common that open standards bodies have spent decades defining exactly what each one means: how a legal entity is identified, how a currency is denoted, how an obligation is expressed, how a payment instruction is structured. TallyUp keeps contract knowledge in those vocabularies — the ones your auditor already trusts — rather than inventing a private one. Ask the record what a party, a price, or an obligation means, and the answer is a definition the wider financial world already agreed on.

We are early, and the boundary is worth stating plainly: the rails proven in production today are bank and ledger — the founder runs his own company on them, every day. The contract surface is where the record extends, and the vocabulary for it is already governed in the graph. That is the direction, stated as a direction.

But the destination is hard to unsee once you’ve seen it. Somewhere in your company’s folder of signed documents sits the answer to the question you’ll be asked next quarter. The only question is whether the record knows it — or whether someone will have to go find the folder again.