Know what happened before the claim arrives.
Every claim decision rests on evidence that was usually collected after the event. Proof builds a verified timeline for every insured field as the season happens: what was planted, how it was managed, what condition it was in. When the claim arrives, the before picture already exists.
A damaged crop is visible. The chain of causation is not.
Was the loss caused by the insured storm, or by disease, a missed application, equipment failure or a condition that predated the event? Every link below has to hold, and each one is usually missing something.
Pre-loss condition
Usually missing entirely. No verified before picture.
Damaging event
Weather data exists, but not tied to the field or the crop’s stage.
Resulting damage
Photographs after the fact. Correlation, not cause.
Production loss
Expected yield is an argument, not a record.
Financial value
Small baseline changes swing the payout by thousands.
Proof does not photograph the damage better. It makes the four links before the damage already exist, dated, when the claim arrives.
Pick a claim. Read the season the way the record kept it.
Three claims from the demo book. Each timeline was built by the farm during the season, for its own reasons, with timestamps that cannot be backfilled. The claim decision reads the timeline instead of arguing with memory.
Loss attribution · from the record
The first evidence you see is a photograph of damage. What the crop looked like three weeks earlier, how it was managed, what it was on course to yield: nobody recorded it, or everybody recorded it differently.
Without a baseline, you are measuring change against an argument.
A satellite shows vegetation declining after a dry spell. It does not show whether drought was the covered cause, or whether pests, disease or a missed fungicide did the work.
Correlation arrives by API. Causation still needs ground truth.
The farmer needs payment this month to keep operating. Verification takes six weeks and an adjuster who cannot reach every farm after a widespread event.
Speed and certainty pull against each other, and both cost money.
Where the record is current, the argument disappears. So does most of the cost.
Demo book figures by peril and evidence state. The pattern is the product: claims on fields with current records settle fast, cheaply and without dispute, because there is nothing left to argue about.
The same record answers both sides of the trust gap.
The farmer’s pain is real too. “I experienced a genuine loss. Why can’t I prove it?” A farm with current records proves it in a day, and gets paid while the neighbours are still photographing.
Records cannot be backfilled. Timestamps come from the recording terminal at the moment of the event, and corrections append beside the original. A season cannot be rewritten after the storm.
A gap is a named gap, never an accusation. Where a record is missing, the timeline says so plainly. Absence of evidence is handled as absence, not as fraud, and both sides see the same gap.
Event response
Sweep the book in minutes
Draw the storm footprint and see every policy in its path with crop, growth stage and management already on record. The demo book reconstructs a field’s pre-loss state in eleven minutes.
Right claim, right payment
Attribution with evidence behind it
The covered share, the pre-existing share and the honestly unattributable share, each traced to records. Fewer overpayments, fewer wrongly refused claims, fewer disputes that reach anyone’s desk.
Renewals without forms
The record chain is the renewal
A farm with current records renews with no forms at all. Underwriters read granted evidence instead of chasing declarations, and the good risks become visible as what they are.
Proof never builds farm risk scores, for you or anyone. Access is scoped to the policy and the season, logged on both sides, and revocable by the farm. A missing record is a named gap, never an assumed fraud. Records cannot be backfilled or edited to flatter, by either side. Aggregates respect the five-record minimum, always.
Proof draws no verdict. It holds the record.
“Why would the evidence exist before we ask for it?”
Because farms record for their own reasons: buyers, banks, schemes and premiums. The same records serve your claim through a grant the farm approves. You are reading evidence that already had a job, which is exactly why it is credible.
“What about farms that do not record?”
Those claims proceed the way they do today. The difference shows up in the book: recorded farms settle faster and cheaper, and the gap between the two is the business case. What share of a portfolio records today is a fair question for the call, with real numbers.
“Can a season be reconstructed after the event?”
Not in Proof. Timestamps are native to the recording terminal, records are append-only, and corrections sit beside the original with author and date. The impossibility of backfilling is not a feature of the product. It is the product.
Proof is working with a first group of agricultural insurers through the 2026 season. There are no testimonials on this page, because we would rather earn real ones than write our own. Early insurers help define the evidence standard their whole market will end up using.
Bring one disputed claim from last season.
20 minutes. We will rebuild it as a timeline and show you what would have been knowable on the day the claim arrived: the before picture, the event, the attribution, and what was honestly unknowable too.