Capture supplier invoices, match them to purchase orders and receipts, propose coding and route approvals — with tolerances, evaluation, audit trails and human approval before any payment is released.
Aggregate processing quality can look acceptable while a small number of invoice cohorts carry most of the exceptions and nearly all of the payment risk. Nestack reports performance by slice, not only in total.
Slice performance — reported separately, not only in aggregateIllustrative example
Slice
Failure rate
Lift
Lift vs. threshold
Status
Invoices without a purchase order
5.2%
3.5×
Review
Partial deliveries and part-invoices
4.1%
2.7×
Review
New and recently changed suppliers
2.9%
1.9×
Watch
PO-matched repeat suppliers
1.0%
0.7×
Normal
Bar: lift vs. PO-matched baseline · scale 0–4.0× · tick marks the 2.0× review threshold2 of 4 slices over threshold
Evidence-linked improvement
The loop does not end at Learn
Each completed cycle leaves the agent with one more regression invoice and one more re-approved control, which is what the next detection is measured against.
Improvement cycle · five stagesSwitchback — the path turns at Improve and returns at Learn
01Detect
Exceptions, coding corrections or duplicate catches rise in a slice.
02Diagnose
Failure traced to extraction, supplier match, tolerance, coding rule or routing.
03Improve
A tolerance, duplicate-check field or routing rule is re-approved and version-linked.
04Verify
The change is re-run against held-out invoices from the affected slice.
05Learn
That invoice becomes a regression case and the changed control enters the runbook.
Learn → DetectThe return edge. Each changed tolerance and rule joins the next cycle's regression set.
Typical build scope
Twelve workstreams across six weeks
The build scope read against the delivery timeline. Week structure follows the six-week plan — discovery, capture, matching and coding, evaluation, integration, then production validation and handover.
WorkstreamWeek 1Week 2Week 3Week 4Week 5Week 6
01Workflow discovery and automation-boundary definition.
02Capture-channel, source-system and API assessment.
03Coding rules, tax codes and tolerance mapping.
04Invoice capture, extraction and normalisation.
05Supplier identification and PO/receipt matching.
06Confidence scoring and exception routing.
07Approval routing and delegation-of-authority rules.
08Human review and exception workflow.
09Duplicate, tolerance and fraud-signal controls.
10Evaluation suite and regression cases.
11ERP integration and payment-run preparation.
12Observability, deployment and Agent Care handover.
12 workstreams · 6 weeks · bar shows the weeks a workstream is active — several run in parallelFinal 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 tierPilotOne capture channelProductionProduction integrationAdvancedMultiple entities / systems
Introduced at Pilot
Extraction and coding proposals✓✓✓
Human approval✓✓✓
Baseline evaluation✓✓✓
Introduced at Production
PO and receipt matching—✓✓
Duplicate and tolerance controls—✓✓
Approval routing—✓✓
Observability and evaluation—✓✓
Payment-run preparation—✓✓
Introduced at Advanced
Multi-entity and multi-currency——✓
Multi-stage approvals——✓
Enterprise controls——✓
Build priceFrom $5,000From $8,000Custom quote
Final build priceConfirmed after discovery based on capture channels, ERP and purchase-order integrations, invoice 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
01Chart of accounts, cost centres and tax codes→Coding-rule, tax-code and tolerance mappingWeek 1
02A representative sample of supplier invoices→Invoice capture, extraction and normalisation, and the matching baselineWeek 2
03Access to relevant APIs, capture channels or exports→Capture-channel and source-system assessment, then integration setupWeek 2
04Supplier master, purchase-order and goods-receipt data→Supplier identification and PO/receipt matchingWeek 3
05AP policy, approval thresholds and delegation of authority→Automation-boundary definition, approval routing and exception routingWeek 3
06Examples of partial deliveries, credit notes and disputes→Evaluation suite, regression cases and failure-mode testingWeek 4
07Named reviewers and payment approvers→Human review workflow, then pilot workflow and production validationWeeks 5–6
Nothing else is requiredDeployment, 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 evaluation and the first parallel run.
PhaseW1W2W3W4W5W6
DiscoveryW1
BuildW2 – W3
EvaluateW4 – W5
Pilot & LaunchW5 – W6
Week focusW1Workflow discovery, AP-policy mapping and automation boundaryW2Capture, extraction and the supplier and PO matching baselineW3Agent workflow, confidence logic and approval routingW4Evaluation suite, duplicate and tolerance controls, failure-mode testingW5ERP integration, pilot workflow and targeted correctionsW6Parallel run against a live payment cycle, then Agent Care handover
Reading the bandBars span only the weeks their work is named in. Integration and evaluation genuinely overlap in week 5.
At the end of W6One payment cycle has run in parallel and been verified, then Agent Care takes over monitoring.
DurationSix-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.
Next step · Accounting AI agent
Build an invoice and AP agent around your payables workflow.
Show us your capture channels, purchase-order and coding rules, and your approval and payment-run controls. We'll map the workflow, identify the automation boundary and recommend the safest path to production.