Nestack Agent Care
Industries / Agriculture / Elevator back office

Agriculture AI agent · Grain handling

Grain-Elevator Back-Office AI Agent (Tickets & Settlements)

Match scale tickets to the producer and contract, apply the posted schedule, do the shrink and discount arithmetic, and draft the settlement — a person approves every payment and receipt.

4–6 weeksTypical delivery
Your stackDeployment
Your sign-offSettlement
Agent CareAfter launch

What this agent does

Does the paperwork, not the payment

In
01

Take the scale tickets as they come off the pit — inbound, outbound and transfer, with weights and grade factors.

02

Take the contracts, the open-storage and receipt records, and the grade and discount schedule posted for that day.

Reason
03

Match each ticket to one producer, one commodity and one contract, and leave the ones that match nothing alone.

04

Work the shrink, moisture and discount arithmetic on the schedule version in force when the load was weighed.

05

Net the position — bushels in, out, on open storage and under receipt — against what the last measurement found.

Decide
06

Separate what the records settle from what they do not, and mark each figure as weighed, graded or calculated.

07

Hold anything that would move a producer's grain, price or balance until that producer and your office have said so.

Out
08

Draft the settlement, the producer statement and the position report with every ticket, schedule and date behind them.

09

Re-check what moved after the draft — late tickets, corrected grades, transfers, cancelled receipts — and flag it.

Product statement

The agent matches, calculates and packages. Every settlement, every payment and every warehouse receipt is approved by a person before it stands.

Example workflow

One scale ticket, end to end

AgentHuman
1Ticket receivedScale ticket from the pit, the probe stand or the outbound loadout
2Producer and contract matchedAgainst the names on the ticket, the delivery record and every open contract for that commodity
3Grade and schedule appliedMoisture, test weight, damage and foreign material against the schedule posted on the weigh date
4Arithmetic drafted and checkedShrink, drying, storage, discounts and deductions, each its own line, each recomputed a second way
No human action required

Stages 1 to 4 run without a person in the loop — the matching and the arithmetic finish before anyone is asked to sign anything. A ticket that matches nothing stops there.

5DecisionSplits on the match and the arithmetic check
Matched, arithmetic agrees

Reaches your office to sign — nothing is written first.

No match, or a figure that fails

Held with the ticket — unsettled and unpaid.

Grain accountant or manager

Reads the ticket, the contract, the schedule version and the arithmetic, then approves, corrects or holds it.

Approve · Correct · Hold
Approved — handed back
6Settlement packagedWritten to the grain system only where an approval has already been recorded against it
7Outcome evaluatedMatch accuracy, arithmetic checks, schedule version, position drift and corrections by producer
Corrections

Every figure your office changes is counted in the evaluation.

What should not run autonomously

Human approval stays in control

Outside the boundary — human approval required8 items
Approving a settlement or releasing a payment.
Issuing, transferring or cancelling a warehouse receipt.
Committing a ticket to a contract or a producer.
Converting open storage into a sale or a price.
Automation boundaryAgent acts unaided
Match each ticket to a producer, a commodity and a contract, and propose it.
Calculate shrink, drying, storage and discounts on the schedule posted that day.
Net the position from tickets, transfers, receipts and the last measurement.
Package the settlement with every figure behind it, and hold what does not match.
Write actions run only inside the approval boundaries agreed in implementation. Paying a producer is not one of them.
Settling a grade dispute or overriding a grade.
Changing the posted grade or discount schedule.
Adjusting inventory to close a position difference.
Changing the match rules, tolerances or thresholds.

Example output

One ticket, annotated

Everything the agent proposes is attached to the ticket and the contract it was matched against.

Settlement output · single inbound loadIllustrative example
Ticket
Delivered by
Weighed
Proposed match
Confidence
Settlement
Inbound corn, one load
Hauler, two producers named
17.4% moisture
Two open contracts, neither exact
88%
Drafted, unapproved, unpaid
As receivedThe ticket as the scale printed it and the names it carries — nothing on this side is recalculated or tidied up.
Evidence used Ticket and weigh date Schedule posted that day Open contract balances
Why it stopped hereTwo contracts are open for this commodity and the ticket names two producers.
ActionApproveCorrectHold
What the score decidesConfidence decides how hard your office checks the match, not whether a producer is paid.

Value

Where AI adds value

The same four claims, placed at the point in the workflow where each one applies.

Where the value landsValue 01 – 04
Every scale ticketInbound, outbound and transfer
03Match & arithmetic

Apply the schedule that was posted

Use the grade and discount schedule, the storage and drying charges and the contract terms in force on the weigh date.

01Approved path

Keep settlement level with the pit

Matching and arithmetic keep up as the loads come in, so what is still on the counter at the end of harvest is what genuinely did not match.

02Human review

Get the exceptions seen the same day

A ticket that matches nothing, a split that needs the producer's instruction and a figure that fails its check reach your office while the driver is still in the yard.

04Build an evidence trail

Retain the ticket, the contract applied, the schedule version, the arithmetic, the approval and the correction — on both paths.

Integrations

Typical integrations

Five system groups connect to the same agent. Which of them are in scope is decided in discovery.

Grain accountingAGRIS · AgVantage · Cultura
Bushel · grain-accounting APIs
Scales & probe standScale indicators · ticket printers
Moisture meters · grading equipment
Contracts & storageContract records · open storage
Warehouse receipts · EWR providers

Agent

Grain-elevator back office

Matches tickets
Calculates settlements
Holds exceptions

Producer paymentsProducer statements · settlement drafts
ACH and cheque runs · deduction records
Observability & evaluationOpenTelemetry · Langfuse
Supported monitoring/evaluation sources

Integration availability depends on the client's existing systems and API access.

Agent controls

Six layers between the model and the settlement

Each control wraps the one inside it. A ticket clears every layer before a settlement is drafted, and the signature sits outside all six.

L6 · Outermost — last line of defenceInward → L1 · closest to the model
L6Rollback / safe modeReturn matching and drafting to your office if signals degrade.Roll back
L5TraceabilityRecord ticket, contract, schedule version, arithmetic, approval and correction.Record
L4Approval gateSettlements, payments and receipts wait for a named person to sign.Gate
L3Schedule pinningOnly the schedule posted on the weigh date can be applied to that load.Pin
L2Second-pass arithmeticShrink, discounts and net value are recomputed and compared.Recompute
L1Match confidenceA ticket that does not resolve to one producer and one contract is held.Hold
Model coreMatch and arithmetic proposed — ticket to contract, grade applied, shrink, discounts and confidence
L1 – L2Decide whether the figures may stand
L3Decides which schedule the load is priced on
L4 – L5Keep the signature with a person and the trail intact
L6Pulls automation back when signals degrade

How Nestack evaluates it

Evaluate the whole ticket-to-settlement path — not only the final figure.

Coverage runs the whole depth of the workflow, and every layer is cut by slice.

Surface — the settlement your office opens
Depth of coverage ▼
E1Final-output evaluationWas the settlement right, line by line, on an independent recalculation?
E2Step-level evaluationWas the ticket matched to the right producer, commodity and contract?
E3Schedule-version checkWas the schedule in force on the weigh date the one that was applied?
E4Tool evaluationDid it read and write the correct ticket, contract and storage record?
E5Slice evaluationHow does accuracy change across producers, commodities and periods?
E6Business outcomeHow much was corrected, re-settled, or found short against the bins?
Floor — the money a producer is actually paid

Failure modes

Where each failure originates in the agent

Seven failure modes plotted against the five stages of the agent lifecycle.

Agent lifecycleDirection of processing →
01 · Ticket intake1 mode
GE-01

Ticket on the wrong producer

A hauled or split load settles to one name.

Stage gathersWeights, grade factors, names and the weigh date
02 · Matching2 modes
GE-02

Wrong contract applied

A settlement drafted on bushels already fully applied.

GE-03

Open storage priced unasked

Stored grain converted to a sale with no instruction.

Stage appliesThe producer, the commodity and the open contract
03 · Calculation1 mode
GE-04

Shrink or moisture arithmetic wrong

Net bushels and the producer's payment are both wrong.

Stage computesShrink, moisture, discounts, storage and drying
04 · Draft / write2 modes
GE-05

Receipt issued or cancelled wrongly

A receipt stands against grain that is not there.

GE-06

Position drifts from the bin

Missed tickets or transfers leave the book overstated.

Stage packagesThe settlement your office reads before signing
05 · Change / Version1 mode
GE-07

Superseded schedule applied

Last week's discounts priced against this week's load.

Stage tracksSchedule, contract-rule and configuration changes
Sev-1 · a producer's money moves unasked Sev-2 · the settlement figure is wrong Sev-3 · the book drifts from the bins

Affected slices

The errors do not spread evenly across the driveway

A few kinds of load carry most of the corrections — split tickets, stored bushels, the days the pit never stops — and they are usually the ones nobody reconciles closely. Nestack reports performance by slice, not only in total.

Slice performance — reported separately, not only in aggregateIllustrative example
SliceFailure rateLift Lift vs. thresholdStatus
Split and hauled tickets7.9%3.6× Review
Harvest-peak delivery days6.0%2.7× Review
Open-storage and price-later4.0%1.8× Watch
Routine single-contract loads2.0%0.9× Normal
Bar: settlement-discrepancy rate lift vs. routine-load baseline · scale 0–4.0× · tick marks the 2.0× threshold 2 of 4 slices over threshold

Evidence-linked improvement

A settlement is corrected in the producer's account, not in the model

Repaying one producer closes that ticket and nothing else. The rule, the schedule handling or the match that produced it is what gets changed and re-tested.

Improvement cycle · five stagesSwitchback — the path turns at Improve and returns at Learn
01Detect

Corrections, re-settlements and position drift show up producer by producer.

02Diagnose

The difference is located in the match, the schedule version, the arithmetic or the source record.

03Improve

Your office re-approves the corrected rule, and the version it takes effect from is written down.

04Verify

Held-out tickets are re-settled and the figures compared with what your office paid.

05Learn

The ticket is re-settled on every release, and the settlement runbook records the rule.

Learn → DetectThe return edge. Telling the producer and putting the money right is your office's job — the agent's part is that the next load is not wrong the same way.

Typical build scope

Twelve workstreams across six weeks

The build scope read against the delivery timeline. Week structure follows the six-week plan — discovery, ticket data, matching and arithmetic, evaluation, integration, then production validation and handover.

Workstream Week 1Week 2Week 3Week 4Week 5Week 6
01Workflow discovery and automation-boundary definition.
02Ticket, contract and storage sources.
03Grade and discount schedule versioning.
04Scale-ticket ingestion and normalisation.
05Producer, commodity and contract matching.
06Shrink, discount and settlement arithmetic.
07Second-pass recalculation and tolerances.
08Approval workflow for settlements.
09Evaluation suite and regression tickets.
10Position against measured inventory.
11Grain-accounting and payment integration.
12Observability, deployment and Agent Care handover.
12 workstreams · 6 weeks · bar shows the weeks a workstream is active — several run in parallel Final scope and sequence confirmed in discovery

Engagement tiers

What each tier includes

Rows are the capabilities named in each tier's scope. Higher tiers include everything below them.

Capability✓ in scope · — not at this tier PilotOne commodity, one location ProductionProduction grain-system integration AdvancedMulti-location / multi-commodity
Introduced at Pilot
Ticket-to-contract matching
Shrink and discount calculation
Settlement drafts for approval
Baseline evaluation
Introduced at Production
Your posted schedules and versioning
Producer statements prepared for review
Approval workflow and audit trail
Observability and evaluation
Introduced at Advanced
Position reconciliation against measurement
Multi-location and multi-commodity books
Enterprise controls and segregation of duties
Build price From $5,000 From $8,000 Custom quote
Final build priceConfirmed after discovery based on locations, commodities, grain-system and payment integrations, ticket volume, approval controls and deployment requirements.
Separate from buildBuild pricing is separate from recurring Agent Care, which covers managed monitoring, evaluations, incidents and verified improvements after launch.

What we need from you

What you bring, and what we build with it

Each input maps to a piece of build scope and a week in the delivery timeline.

You bringWe build with it
01Your posted grade and discount schedules, with the dates they applied Schedule versioning and the rule that picks a load's scheduleWeek 1
02A season of scale tickets, including the ones that had to be redone Ticket ingestion, normalisation and the matching baselineWeek 2
03Contract, open-storage and warehouse-receipt records Producer, commodity and contract matchingWeek 3
04How you calculate shrink, drying and storage charges today Settlement arithmetic and the second-pass recalculationWeek 3
05Who may approve a settlement, a payment and a receipt Approval workflow, audit trail and the automation boundaryWeek 4
06Your last position reconciliation and the bin measurements behind it Position reconciliation and the drift check against the bookWeek 5
07Access to the grain accounting and payment systems Source assessment, then grain-system and payment integrationWeeks 5–6
Nothing else is required Deployment, documentation and Agent Care handover are ours.

Delivery timeline

Four phases across six weeks

Phases are drawn over the weeks they actually occupy. Week 5 carries both the position reconciliation and the first settlements drafted from live tickets.

Phase W1W2W3W4W5W6
Discovery W1
Build W2 – W3
Evaluate W4 – W5
Pilot & Launch W5 – W6
Week focus W1Workflow discovery, schedule versioning and the approval boundary W2Scale-ticket ingestion, normalisation and the matching baseline W3Matching, settlement arithmetic and the second-pass check W4Evaluation suite, approval workflow and failure-mode testing W5Integration, position reconciliation and drafting from live tickets W6Your office settles from drafted tickets, then Agent Care starts
Reading the bandThe arithmetic is tested against your own settled tickets in week 4, before anything is drafted for approval in week 5.
At the end of W6Settlements have run alongside your existing process and been approved by the people who sign them, then Agent Care takes over monitoring.
DurationSix-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.

Next step · Agriculture AI agent

Build a back-office agent around your ticket and settlement work.

Show us a week of scale tickets, the schedules you posted against them and the contracts they should have landed on. We'll settle that week from your own records and show you every ticket and every figure ours disagrees with.

Nestack Agents · Grain-elevator back officeAGT-AGR-12 · Agent Care available after launch