Systems and records
Agriculture contract settlement data: linking statements to evidence
By SourceX Editorial · Reviewed by Noah Loul ·
9 sections
Short answer
Link agricultural settlement statements to evidence line by line: each line should point to the delivery ticket and lot it pays for, the quality result in force on that date, the contract clause and schedule that turned the grade into an adjustment, any approval for a substitute or exception, and the payment. A line that recomputes from those records is reconciled; one that does not is the task.
License terms
Agriculture contract
Key takeaways
- A settlement statement summarizes decisions; the contract, delivery and quality records show how each was made.
- Reconcile line by line: every adjustment should recompute from a quality result and a contract schedule in force on the delivery date.
- Keep original and corrected quality results as separate, linked events so the task can be replayed.
- A paid status is not a reconciled settlement, especially when part of the amount is disputed.
- Include only contracts and records the business is permitted to use, and treat ranking and pricing data with care.
How do you link agricultural settlement statements with contract and quality evidence?
Link them line by line through shared keys: each settlement line should point to the delivery it pays for (scale or receiving ticket and lot), the quality result that set its grade or adjustment (certificate, inspection or lab test, with its version), the contract clause and schedule in force on the delivery date, any approval for a substitute or exception, and the payment that settled it. When every line recomputes from those records, the statement is reconciled; when one does not, the gap is the task.
Agriculture contract settlement data is valuable to AI teams for that reason. A settlement statement on its own is a summary of decisions already made. Joined to the contract and the quality record, it becomes a sequence a model can learn from: what was delivered, how it was graded, which rule turned the grade into a payment adjustment, who approved a deviation, and whether the final payment matched.
The asset on this page is a task pack built for that join, labeled permitted because only contracts and records the business is allowed to use belong in it. The examples are illustrative, and the regulatory references are United States rules that shape what records exist, not advice on any specific contract.
What does a settlement statement rest on in three common contract types?
The documents differ by commodity and contract type, but each settlement depends on a delivery record, a quality determination and a contract term. Three common United States arrangements show the pattern, and several of them have record rules that tell you what should exist.
| Arrangement | Settlement document | Quality evidence | Record rule worth knowing |
|---|---|---|---|
| Poultry grower contract with a live poultry dealer | Settlement sheet (final accounting) and, where pay depends on ranking, a ranking sheet | USDA condemnation or grade certificates when they affect pay | 9 CFR 201.100 requires an accurate settlement sheet at settlement and the certificates at or before settlement |
| Grain delivered to an elevator or processor | Settlement or purchase statement built from scale tickets | Elevator grading or an official FGIS certificate reporting grade and factors such as test weight, moisture and damage | Official certificates are issued under the U.S. Grain Standards Act for grains with official standards |
| Fresh produce sold under PACA | Invoice or, on consignment, an account of sale | Receiving notes and, in disputes, an inspection certificate | 7 CFR 46.2 sets prompt payment periods and limits them to the undisputed amount when a dispute exists |
| Any of the above, for listed foods | Shipping and receiving records | Traceability lot code and key data elements | FDA's food traceability rule ties key data elements to critical tracking events |
What belongs in a permitted reconciliation pack for one settlement period?
Build one pack per settlement period per counterparty, or per disputed line when the work is narrower. Fill every element, write none where nothing applies, and give each a reference to a record a reviewer can open. Mark each pack real or illustrative and record its version.
| Element | Contents | Source record | Reconciliation check |
|---|---|---|---|
| Permission scope | Which contract, counterparty class and period may be used; confidentiality limits | Contract confidentiality clause, internal approval to use the records | Pack excluded if permission is unclear |
| Contract version | Price basis, premium and discount schedule, delivery period, substitution terms | Executed contract and amendments with effective dates | Version in force on each delivery date |
| Delivery events | Ticket or receipt number, date, lot or load, quantity received | Scale tickets, receiving records, shipping records | Quantities sum to the statement's quantities |
| Quality results | Grade, factors, test or inspection IDs, result version and time | Certificates, inspection reports, lab results | Each adjustment traces to a result that existed before settlement |
| Exceptions and approvals | Rejections, substitute lots, regrades, waived terms and who approved | Approval log, notices, correspondence | Every deviation from contract has an approval |
| Settlement lines | Each line's basis, quantity, grade adjustment and deductions as stated | Settlement statement or account of sale | Each line recomputes from the records above |
| Payment and disputes | Payment date, amount reference, disputed and undisputed portions | Remittance, payment record, dispute file | Undisputed portion paid within the agreed or regulatory period |
| Traceability links | Lot codes and event IDs linking delivery, quality and onward shipment | Traceability records or EPCIS events | Chain replays without gaps |
| Verified outcome | Reconciled, reconciled after correction, or unresolved | Pack review record | Outcome label backed by the checks |
How can lot codes and corrected quality results make a settlement replayable?
A settlement can be replayed if the pack keeps event identifiers and corrections instead of only final values. A replayable task lets a reviewer, or a model, walk from the delivery to the quality decision to the payment in time order and see what was known at each step.
For foods on FDA's Food Traceability List, the food traceability rule requires covered firms to keep key data elements tied to critical tracking events such as harvesting, cooling, initial packing, shipping, receiving and transformation, linked to a traceability lot code. FDA says it will not enforce the rule before July 20, 2028, but many supply chains already record these elements, and they give the settlement pack its lot spine.
GS1's EPCIS standard offers a neutral event format for the same chain. Its JSON schema lets an event carry business transaction references typed as purchase order, invoice, receiving advice, certificate or test result, which is exactly the set of documents a settlement reconciliation joins. It also defines an error declaration with a declaration time, a reason and corrective event IDs, and the companion vocabulary defines the incorrect data reason as an event whose data were wrong and which later events may correct. A regrade that replaces an earlier result fits that model, and keeping both events is what makes the task replayable.
- Keep the original quality result and the regrade as separate events, linked by a correction reference.
- Store the time each result became available, so a reviewer can check whether settlement used the result in force.
- Carry the lot code through substitutions so the substitute lot's quality record joins the right settlement line.
- Record onward shipments when a rejected lot was diverted, since diversion often explains a later adjustment.
Which record proves each settlement decision?
Each kind of decision has one record that proves it, and the pack should name that record rather than relying on the statement's narrative. The table pairs common settlement decisions with the evidence that settles them and the outcome check that closes the task.
| Decision on the statement | Proving record | Outcome check | Label if the record is missing |
|---|---|---|---|
| Grade discount or premium | Quality result in force on the delivery date | Adjustment recomputes from the contract schedule | Unsupported adjustment |
| Rejected or partly accepted load | Receiving inspection and rejection notice | Quantity paid matches quantity accepted | Unsupported rejection |
| Substitute lot accepted | Substitution approval under the contract terms | Substitute lot's quality joins the line | Unapproved substitution |
| Ranking-based pay | Ranking sheet with the figures used | Grower's position recomputes from the figures | Ranking not reproducible |
| Deduction for a dispute | Dispute record separating disputed and undisputed portions | Undisputed portion paid on time | Dispute handling unclear |
| Final payment | Remittance or payment record | Paid amount equals reconciled total | Payment not verified |
How would an illustrative downgraded load with a substitute lot be reconciled?
This case is invented; every count and percentage is illustrative and no real contract is described. A vegetable packer buys from a grower under a seasonal contract with a quality schedule and a clause allowing a substitute lot with the buyer's written approval. Over one week the grower delivers three loads.
Load two fails a receiving inspection for size grade, with about 12 percent of units below specification. The packer's quality team records the result, and the grower asks to replace part of the load with a lot from another field. A buyer approves the substitute by email the same afternoon, and the substitute arrives the next day under its own lot code with its own receiving record.
The weekly settlement statement shows load two with a size discount applied to the full original quantity and the substitute lot paid at full grade. The pack's checks find two problems. The discount was applied to units that were returned and replaced, so the quantity basis is wrong, and the substitute lot's quality result was updated by a re-inspection two days after settlement, which the statement did not use.
The outcome is reconciled after correction: a corrected statement applies the discount only to the units retained and uses the re-inspection as a correction of the first result, with both results kept. The task a model learns from this is not settle the statement but find which record each line depends on, notice when a later record supersedes an earlier one, and correct only what the evidence supports.
Why can a paid-in-full status still leave a settlement unresolved?
A payment status of paid in full does not mean the settlement reconciled. In a variant of the same case, the packer pays the statement as issued and the grower raises a dispute about the discount. The payment system shows the invoice closed, and a model trained on statuses would learn that the case ended well.
Under PACA's prompt payment rules, when a dispute exists the payment periods apply to the undisputed amount, so a payment that went out on time can still leave a disputed portion open. The pack labels this case unresolved until the dispute record shows how the disputed portion was settled. A plausible closing message in the email thread, such as all sorted, thanks, is not evidence; the dispute file and the final remittance are.
A second failure pattern is a grade on the statement with no quality record behind it, often because the grade was typed from a phone call. The pack marks that line unsupported rather than guessing, and the case becomes a training example of what to ask for, not of how the line was paid.
Which contracts and records are permitted in the pack?
Only include contracts and records the business is allowed to use for this purpose, and decide that before building packs. Grower and buyer contracts often carry confidentiality clauses, ranking sheets contain other growers' performance figures even when names are hidden, and pricing schedules are commercially sensitive. Each of these needs a decision, not a default.
Keep the distinctions straight. Holding the records, having any privacy permission needed for people named in them, owning the records and holding contractual rights to use them are separate questions, and removing names does not by itself create a right to license. Counsel should review contract terms before any pack leaves the business.
- Replace grower, farm and field identifiers with stable pseudonymous keys, and generalize locations to region.
- Express prices and adjustments as rule references and ratios, or remove them, rather than sharing schedules as written.
- Exclude ranking sheets unless the other growers' figures can be aggregated so no single grower can be inferred.
- Keep individual inspectors and employees as roles.
- Leave out any counterparty whose contract forbids secondary use.
Can reconciled settlement histories be licensed as AI data?
Packers, processors, elevators, integrators and distributors with years of reconciled settlement histories hold examples that are difficult to collect any other way: real quality decisions, real exceptions and payments that can be checked against them. Prepared as permitted task packs, those histories can show an AI team how settlement decisions are actually made.
A packer or elevator that wants to license its settlement packs can work through SourceX, which handles the rights review, the contract and the preparation that make selected records an AI dataset. Finding out whether the business qualifies takes a short assessment and no uploads, and no pack is delivered until the business has signed off on scope and terms and signed the agreement.
Frequently asked questions
What is agriculture contract settlement data?
It is the set of records behind a payment to a grower or supplier under a contract: delivery tickets, quality results, the contract schedule, approvals for exceptions, the settlement statement and the payment. Linked together, they show how each payment line was decided.
Do poultry growers have a right to see the records behind their settlement?
Federal rules at 9 CFR 201.100 require live poultry dealers to give growers an accurate settlement sheet at settlement, a ranking sheet when pay depends on ranking, and relevant USDA condemnation or grade certificates. Counsel should confirm how the current rule applies to a specific contract.
Should a regraded quality result replace the original in the dataset?
No. Keep both, with the regrade recorded as a correction of the original and the time each became available. That is what lets a reviewer check which result the settlement actually used.
How do you handle a settlement that is under dispute?
Separate the disputed and undisputed portions in the pack, check the undisputed portion against the payment terms, and label the case unresolved until the dispute record shows the outcome.
Can settlement records be shared if grower names are removed?
Removing names helps with privacy but does not settle contractual rights. Confidentiality clauses, ranking data about other growers and pricing schedules each need a decision, and counsel should review them first.
Sources
- 9 CFR 201.100 requires live poultry dealers to give growers an accurate settlement sheet at settlement, a ranking sheet showing the figures used when pay depends on ranking, and USDA condemnation or grade certificates that affect pay at or before settlement. Source
- 7 CFR 46.2(aa) defines full payment promptly under PACA, including payment within 10 days after the day produce is accepted, written agreements on other terms, and that in a dispute the periods apply only to the undisputed amount. Source
- FDA's food traceability rule requires covered firms to keep key data elements tied to critical tracking events and traceability lot codes, and FDA will not enforce it before July 20, 2028. Source
- USDA AMS issues official certificates for grains with standards under the U.S. Grain Standards Act, reporting grade based on factors such as test weight, moisture, cleanliness and damage. Source
- The GS1 EPCIS 2.0 JSON schema types business transactions as purchase order, invoice, receiving advice, certificate, test result and others, and defines an error declaration with declaration time, reason and corrective event IDs. Source
- The GS1 Core Business Vocabulary defines the incorrect data error reason, test result and receiving advice transaction types. Source
See if your company qualifies
A short company assessment. No data uploads are needed.