Mining AI agent · Royalty returns
Mineral Royalty Return AI Agent
Assemble the support behind a royalty return — quantities reconciled from weighbridge to invoice, deductions evidenced, the amendment history — for the named officer who lodges. The fiscal filing, not the tenement clock.
- Typical delivery
- 4–6 weeks
- Deployment
- Your stack
- Officer lodges
- Support only
- After launch
- Agent Care
What this agent does
Assembles the support, not the liability
In
- 01
A load crosses the weighbridge, and the record is pulled from supported despatch, shipping and sales sources.
- 02
A wet tonne is carried with the moisture certificate behind it, not converted on a standing factor.
Reason
- 03
A return falls due, and the period is read against the calendar the revenue authority publishes, not a vendor table.
- 04
A deduction is claimed, and the contract or invoice that evidences it is attached to the line it supports.
- 05
A provisional invoice settles, and the price movement is raised against the period it belongs to.
Decide
- 06
A sale goes to a related party, and the line is held until the valuation decision on file is produced.
- 07
A determination governs the base, and the shipment is checked against the product it was made for.
Out
- 08
A period is retained as a dated version, beside the records behind it and what the lodging officer was shown.
- 09
A write action runs only inside the approval boundaries agreed during implementation.
Product statement
The agent assembles and reconciles the support; a named officer lodges the return, and the tenement holder or royalty payer carries the liability.
Example workflow
One period, tonne to lodgement
- 1Records receivedWeighbridge, despatch, shipping and sales-invoice records
- 2Support assembledQuantities, deductions, price adjustments and prior amendments, each with its named source
- 3Return pack draftedReconciled quantities, deduction basis, open items and confidence
- 4Controls appliedQuantity reconciliation, deduction-evidence tests, calendar checks and confidence threshold
No human action required
Stages 1 to 4 run unaided, and nothing is lodged at any of them — the agent is assembling support, and the officer lane opens at the confidence gate.
Goes to the royalty officer.
Adds a commercial read first.
Officer review
The pack is held with its records, its open items and the confidence.
Approve · Amend · Send to commercial review
Approved — released to lodge
- 6Royalty records updatedOnly where write access and approval policy allow it
- 7Outcome evaluatedAmendment distance, open-item outcomes, reassessments and post-lodgement corrections
Amendments
Every officer amendment is counted in the evaluation.
What should not run autonomously
Human approval stays in control
Outside the boundary — human approval required
8 items
- Lodging the return — a named officer does that.
- Determining the royalty liability for a period.
- Deciding whether a deduction is allowable.
- Valuing a related-party or non-market sale.
Automation boundary
Agent acts unaided
- ✓Reconcile production and sales quantities to the invoiced tonnes.
- ✓Carry each deduction claimed with the evidence held on file behind that claim.
- ✓Track provisional-to-final pricing and raise what has moved.
- ✓Assemble the audit pack and hold it for the named officer.
Writes run only inside the boundaries agreed at implementation, and never ahead of the officer.
- Applying for a valuation decision or determination.
- Certifying or verifying the lodged statement.
- Notifying the Commissioner of an understated liability.
- Changes to deduction, calendar or approval rules.
Example output
One return line, annotated
Everything the agent assembles is attached to the record it was drawn from.
Return support output · single line
Illustrative example- Commodity
- Concentrate shipment
- Return line
- Wet despatch tonnes converted on the certificate for this shipment
- Reconciled quantity
- 1,284 dry t
- Source of record
- Moisture certificate
- Confidence
- 88%
- Basis of record
- Weighbridge and shipping records
As received
Taken from the weighbridge record and the moisture certificate — nothing on this side is written by the agent.
Source facts used
- Weighbridge record
- Moisture certificate
- Shipping documents
Why this is open
The invoice is struck on dry tonnes and a person still decides.
What the score decides
Below the configured threshold the line picks up a commercial read before the officer sees it.
Action
ApproveAmendSend to commercial reviewValue
Where AI adds value
The same four claims, placed at the point in the workflow where each one applies.
Where the value lands
Value 01 – 04
Every period
From the despatch and sales records
Reconcile the tonnes to the invoice
Quantities are set against the records already on file; what the return states stays with the officer who lodges it.
Lodging is the assessment
In Queensland, lodging the return makes the assessment — Taxation Administration Act 2001 s.14.
This is not the tenement clock
The shipped tenement agent tracks what a tenement obliges you to do. This page asks whether the lodged number is defensible.
Build an evidence trail
The tonne, the weighbridge record under it and the officer who lodged stay on the return.
Integrations
Typical integrations
Five system groups connect to the same agent. Which of them are in scope is decided in discovery.
Weighbridge and production
- Weighbridge
- scale systems
- Production reporting
Despatch and shipping
- Rail and port despatch
- Bills of lading
- draft survey
Sales and invoicing
- SAP
- Oracle ERP
- Offtake and settlement
Agent
Royalty return support
- Reads the records
- Assembles the pack
- Holds for lodgement
Assay and lodgement
- LIMS
- moisture certificates
- Revenue authority portals
Observability & evaluation
- OpenTelemetry
- 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 return
Six layers, narrowing inward. What survives the last of them is named in the map below.
L6 · Outermost — last line of defence
Inward → L1 · closest to the model
- L1 – L2
- Test whether the pack may stand
- L3
- Puts the release in the hands of an officer
- L4 – L5
- Keep the tonne and the record behind it
- L6
- Holds the return at unlodged when signals degrade
How Nestack evaluates it
Evaluate the assembly workflow — not only the finished return.
Coverage runs the whole depth of the workflow, and every layer is cut by slice.
Surface — the return the officer lodges
- E1Final-output evaluationDid every quantity trace to the record behind it?
- E2Step-level evaluationDid the agent use the right calendar, rates and deduction configuration?
- E3Tool evaluationDid it read and write the correct entity and the correct period?
- E4Confidence calibrationDo low-confidence lines actually attract more officer amendments?
- E5Slice evaluationHow does performance change across specific commodity groups?
- E6Business outcomeHow many lines were amended, or corrected after the return was lodged?
Floor — the outcome the payer answers for
Failure modes
Where each failure originates in the agent
Seven failure modes, each placed where it first enters the return.
Agent lifecycle
Direction of processing →
01 · Retrieval
1 mode
- DX-03
Certificate never ingested
The conversion runs on a standing factor.
Stage gathers
Weighbridge, despatch, shipping and sales records
02 · Reasoning
2 modes
- DX-04
Standard moisture factor
A wet-to-dry conversion runs without its certificate.
- DX-06
Deduction claimed by analogy
A cost is claimed against a rule that does not reach it.
Stage proposes
Reconciled quantities, deductions and confidence
03 · Tool / write
2 modes
- DX-02
Vendor-sourced due date
Instalment dates inside the quarter are missed.
- DX-05
Final pricing never raised
The settlement moves after the period is lodged.
Stage writes
Only where write access and approval policy allow it
04 · Output
1 mode
- DX-01
Quantity basis mismatch
Mined tonnes are returned where disposed tonnes belong.
Stage returns
The return the officer lodges and the auditor reads
05 · Change / Version
1 mode
- DX-07
Silent schedule drift
A rule change alters which rate a product attracts.
Stage tracks
Model, prompt, deduction rules and calendar config
- Sev-1 · a return lodged outside the boundary
- Sev-2 · a wrong tonne reaches the return
- Sev-3 · source degrades, line goes to review
Affected slices
A clean period total can hide one commodity
A period-level reconciliation figure can read as clean while a few commodity and sales-type slices carry most of the returned lines. Nestack reports the return rate by slice, not only in total.
Slice performance — reported separately, not only in aggregate
Illustrative example| Slice | Failure rate | Lift | Lift vs. threshold | Status |
|---|---|---|---|---|
| Related-party concentrate sales | 7.4% | 3.6× | Review | |
| Provisionally priced shipments | 5.1% | 2.5× | Review | |
| Iron ore under determination | 3.8% | 1.9× | Watch | |
| Final-priced domestic sales | 2.5% | 1.2× | Normal |
Bar: return-rate lift vs. domestic final-priced baseline · scale 0–4.0× · tick marks the 2.0× review threshold
2 of 4 slices over threshold
Evidence-linked improvement
Each cycle leaves one more test behind
A cycle closes when the amended return is a regression case. That suite is what the next return lodged is measured against.
Improvement cycle · five stages
Switchback — the path turns at Improve and returns at Learn
- 01
Detect
Return rate rises for one commodity slice.
- 02
Diagnose
The quarter that was lodged on time and re-opened long after the team had moved on is read back until the cause narrows to one.
- 03
Improve
The change goes out with a number, and the returns behind it travel with it.
- 04
Verify
Nothing lodges while one touched return case is still red.
- 05
Learn
It is kept for good, and the deduction rules are amended in the same commit.
Learn → Detect
The return edge. Detection next period starts from a suite one case longer.
Typical build scope
Twelve workstreams across six weeks
The build scope read against the delivery timeline. Week structure follows the six-week plan — discovery, sources, return workflow, evaluation, integration, then production validation and handover.
Workstream
Week 1Week 2Week 3Week 4Week 5Week 6
- 01Royalty workflow discovery and boundary mapping.Week 1
- 02Weighbridge and sales source assessment.Week 1, Week 2
- 03Quantity, deduction and lodgement-calendar mapping.Week 1, Week 2
- 04Record ingestion and reconciliation.Week 2
- 05Pack assembly and source binding.Week 2, Week 3
- 06Confidence scoring and exception routing.Week 3
- 07Officer review workflow.Week 3, Week 4
- 08Weighbridge and sales-system integration.Week 5
- 09Quantity and deduction cases.Week 4, Week 5
- 10Guardrails and lodgement controls.Week 4
- 11Return-trail instrumentation.Week 5, Week 6
- 12Deployment, documentation and Agent Care handover.Week 6
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 period | ProductionProduction source systems | AdvancedMultiple states / entities |
|---|---|---|---|
| Introduced at Pilot | |||
| Return support and reconciliation | ✓ | ✓ | ✓ |
| Officer lodgement approval | ✓ | ✓ | ✓ |
| Quantity-reconciliation baseline | ✓ | ✓ | ✓ |
| Introduced at Production | |||
| Reporting by commodity | — | ✓ | ✓ |
| Exception workflow in your systems | — | ✓ | ✓ |
| Approved write-back | — | ✓ | ✓ |
| Weighbridge-and-sales integration | — | ✓ | ✓ |
| Introduced at Advanced | |||
| Multi-jurisdiction deduction rules | — | — | ✓ |
| Multi-stage commercial approvals | — | — | ✓ |
| High shipment volume | — | — | ✓ |
| Multi-state return controls | — | — | ✓ |
| Build price | From $5,000 | From $8,000 | Custom quote |
- Final build price
- Confirmed after discovery based on integrations, workflow complexity, transaction volume, approval controls and deployment requirements.
- Separate from build
- Build 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 bring
We build with it
01Your commodity list and return periods
→Quantity, deduction and calendar rule mapping
Week 102Representative prior lodged returns
→Record ingestion, reconciliation and the assembly baseline
Week 203Your deduction policy and evidence standards
→Return-support mapping and automation-boundary definition
Week 104Access to relevant APIs, feeds or exports
→Weighbridge and sales-system assessment, then integration setup
Week 205Returns you would not want audited
→Amendment cases and failure-mode testing
Week 406What no lodged return may leave unsupported
→Confidence scoring, exception routing, guardrails and lodgement controls
Week 307Named officers to review the return pack
→Officer review workflow, then pilot and production validation
Weeks 5–6
Nothing else is required
Deployment, documentation and Agent Care handover are ours.
Delivery timeline
Four phases across six weeks
Bands here are measured in weeks of real work, and that measurement is what puts two into week five.
Phase
W1
W2
W3
W4
W5
W6
- DiscoveryW1
- BuildW2 – W3
- EvaluateW4 – W5
- Pilot & LaunchW5 – W6
Week focus
W1Royalty workflow discovery, rule mapping and the automation boundary
W2Source integration and the reconciliation baseline
W3Assembly workflow, confidence logic and lodgement controls
W4Evaluation suite, deduction checks and failure-mode testing
W5Sales-system integration, a pilot period and targeted corrections
W6One return period run under the commercial manager, then Agent Care handover
- Reading the band
- A bar covers only the weeks its own work is named in. The week 5 overlap is real, not padding.
- At the end of W6
- When the return period validates, Agent Care takes the agent on.
- Duration
- Six-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.
Next step · Mining AI agent
Build the support behind the return before the period closes.
Show us your despatch records, your deduction basis and who lodges. The commercial manager who presses submit is the person an auditor will ask, years later, what each tonne on that return was drawn from.
- Typical delivery
- 4–6 weeks
- Deployment
- Your production and sales data
- Officer lodges
- Support only
- After launch
- Agent Care available
Nestack Agents · Mineral royalty returns
AGT-MN-15 · Agent Care available after launch