Why Contract Data Alone Isn't Enough

Contract intelligence tools have become very good at finding clauses. Give one a transportation agreement and it will locate the detention provision, the free-time allowance, the rate and the cap. This is genuinely useful work, and it is a prerequisite for everything that follows. It is also, on its own, not verification.

A clause is a conditional statement

"Carrier shall be allowed 120 minutes of free time per stop; detention accrues thereafter at $75.00 per hour in 15-minute increments." Read carefully, this sentence tells you nothing about what is owed. It tells you what would be owed if a certain state of the world obtained. The antecedent — dwell exceeded 120 minutes — is an empirical claim about a physical event.

Comparing a clause to an invoice therefore answers a narrower question than it appears to: given the biller's own account of the facts, is the amount internally consistent? That check will catch a rate error or a bad increment. It cannot catch a wrong duration, because both documents in the comparison come from the same side.

Contract data tells you the rule. Evidence tells you whether the rule was triggered.

Three failure modes of contract-only review

• Unverified antecedents — the condition is assumed rather than measured.
• Ambiguous measurement points — the clause says "arrival" and the reviewer silently picks whichever timestamp is available.
• Silent completeness assumptions — a charge is approved because nothing contradicted it, not because anything supported it.

What to add

The missing layer has four parts. First, obligations as structured objects: free time in minutes, rate in cents, increment, cap, and the specific measurement points the clause names. Second, an explicit statement of which evidence types could establish those measurement points. Third, event reconstruction from independent sources, with disagreement surfaced rather than averaged away. Fourth, a deterministic calculation that is reproducible from the recorded inputs.

The payoff is that "unsupported" and "insufficient evidence" become different answers. A charge with no supporting evidence is not the same as a charge the evidence refutes, and treating them identically is how disputes get lost.