How long does deployment typically take?

Most customers go live in 8–12 weeks; multi-country, multi-carrier, regulated-industry deployments run 14–16 weeks. The pilot carrier and country are typically live in weeks 6–9 with parallel-run audit.

How does Nexa handle contracts that are PDF-only or informally structured?

We digitise contracts into a machine-readable object store as part of deployment — rate matrix, accessorial rules, fuel linkage, discount tiers, volume thresholds. For informal contracts, we build the object during discovery workshops with procurement. Once digitised, contracts are versioned and auditable.

What happens when Nexa can't match a line?

She packages the evidence — the invoice line, the expected rate from contract, the shipment actuals, the POD status — and routes to Vera for dispute resolution. Nothing gets silently paid or silently rejected. Every exception has a trail.

Can Nexa handle both inbound freight invoices and outbound 3PL payouts?

Yes. The matching engine is bidirectional. On the inbound side she audits what carriers bill us; on the outbound side she computes what we owe drivers, 3PLs, and middle-mile partners based on shipment events and contract terms. Both close back into the ERP.

How do you prevent duplicate payments or fraudulent invoices?

Nexa runs duplicate detection on every invoice at ingestion — invoice number, amount, date, carrier, shipment reference. She also runs anomaly detection across carrier behaviour (unusually high accessorial rates, repeated round-number variances, ghost shipment patterns). Suspicious lines are always routed to human review.

How does Nexa integrate with our ERP and close process?

Standard connectors to Oracle ERP, SAP, NetSuite, Dynamics. Released lines post as approved invoices with remittance advice; disputed lines hold in a suspense account until Vera resolves. Month-end close on freight cost typically moves from 30–45 days to 5–10 days.

What about multi-currency and cross-border?

Nexa handles multi-currency natively — FX captured at invoice-date rate, settled in the entity's functional currency, consolidated at group level per policy. Country-specific tax and e-invoicing rules (IRN, Peppol, GSTN) are handled by dedicated connectors.

How is the leakage-recovered value typically measured?

During pilot, we run Nexa in shadow mode on 30–60 days of historical invoices and surface the cumulative variance Nexa would have caught. That number — typically 1.5–3.5% of freight spend — becomes the baseline leakage-recovered target for steady state.

How does Nexa work with our internal audit and SOX controls?

Segregation-of-duty is enforced at the engine level — the matching engine cannot trigger a payout without a policy-approved release. The audit log is immutable. Internal audit teams are given read-only access to the full decision trail, the contract versions, and the tolerance policies. Annual SOX testing has passed at enterprise customers operating in listed-company reporting environments.

What happens on Day 1 at go-live vs. steady state?

Day 1: Nexa processes every new invoice from the cut-over carrier in parallel with the existing workflow. Humans review every release for the first two weeks. As accuracy is confirmed, the human-review rate drops (typically to under 20% by week 4, under 5% by week 8). At steady state, human attention is reserved for true exceptions — new contract structures, atypical accessorial patterns, new carriers.

Can Nexa handle volume-based discounts and end-of-period true-ups?

Yes. Volume discount tiers are part of the contract object. Nexa tracks running volumes per carrier per contract period and flags when a tier transition should have been applied. End-of-period true-ups are computed and posted as adjustments, not left to quarterly manual reconciliation.

Nexa — The Settlement Agent

The problem it solves

Three percent of every freight invoice leaks. The rate card on the PDF doesn’t match the contract, the accessorial charges are applied the wrong way, the POD exists but isn’t linked to the invoice line, the fuel surcharge is calculated against the wrong week’s index. Finance teams know the leakage is real — they just can’t catch it at scale. So they either wave invoices through and rebuild margin somewhere else, or they staff a 12-person audit team that closes the books 45 days after month-end.

The second pain is the outbound mirror of the first: paying drivers, 3PLs, and carriers correctly and on time. Manual reconciliation of POD-to-invoice, performance-based bonuses, credit notes for failed deliveries, multi-currency settlement across country entities — all of it sitting in spreadsheets, Excel errors creeping in, vendors chasing payments, working capital trapped.

What it is

Nexa is the Settlement Agent inside AgentFleet — the component that owns end-to-end freight audit, rate-card enforcement, POD reconciliation, and vendor payout. She’s the autonomous counterpart to a freight audit-and-pay team: she ingests every inbound invoice, matches every line to contract, validates every POD, flags every exception, and either releases for payment or routes to dispute with the evidence packet pre-built.

Nexa is new-age because she operates on structured contract objects, not PDFs. Rate cards live as machine-readable objects; accessorial rules are versioned; fuel indexes are pulled from primary sources; POD data comes from the same driver app that captured the delivery. The result is a closed loop: a settlement decision that ties back to a specific shipment event, a specific contract clause, and a specific POD — with the reasoning exposed. When Nexa can’t settle autonomously, she hands off to Vera with a fully-built dispute case.

Core capabilities

Capability	What it does
Invoice ingestion & normalisation	Ingests invoices via EDI, email PDF, API, portal scraping. Parses line items into a canonical schema regardless of carrier format.
Contract & rate-card object store	Contracts live as structured objects — base rate, zone matrix, weight breaks, accessorials, fuel index linkage, volume discount tiers — versioned and auditable.
Line-by-line match	Every invoice line matched against expected rate for that shipment: origin, destination, weight, dim, service level, date, carrier.
POD validation	Each billed shipment must have a matched POD (ePOD, signature, OTP, photo) before release. Missing or mismatched PODs auto-fail the line.
Accessorial audit	Detention, waiting time, re-attempt, COD collection, insurance — validated against the contractual rule, not the carrier’s claim.
Multi-currency & tax handling	Multi-country, multi-currency, multi-tax-jurisdiction (VAT, GST, IGST, sales tax) with automated FX capture at invoice-date rate.
Three-way match (PO–invoice–POD)	Where a purchase order exists (primary freight), Nexa runs the full three-way match before release.
Duplicate & fraud detection	Pattern matching across invoices for duplicate submissions, round-tripping, ghost shipments, unusually high accessorials — flagged before payment.
Dispute case packaging	When a line fails, Nexa assembles the evidence — contract clause, actual shipment event, POD, carrier claim — and hands to Vera for resolution.
Carrier payout orchestration	Once approved, Nexa triggers payout through the ERP or treasury system with full remittance advice.
Performance-based scorecards	Settlement data feeds a carrier performance layer — on-time %, POD completion %, claim rate — used by Multi-Carrier Intelligence for future allocation.
Close-the-books acceleration	Month-end close on freight cost typically moves from 30–45 days to 5–10 days after go-live.

How it works

Nexa sits between the carrier invoice stream and the ERP payout workflow. Every inbound invoice runs through a sequence — ingest, normalise, match against contract, validate POD, compute variance, decide action. The contract store is the source of truth; the shipment event stream from Atlas is the source of actuals.

graph TB A[Carrier invoices
EDI, PDF, API, portal] --> B[Ingestion & parsing] B --> C[Canonical invoice schema] D[Contract object store
rate cards, accessorials, fuel index] --> E[Matching engine] F[Shipment event store
Atlas] --> E G[POD repository
ePOD, signatures, OTP] --> E C --> E E --> H{Variance engine} H -->|within tolerance| I[Auto-release to ERP payout] H -->|exception| J[Vera — dispute case] I --> K[Carrier scorecard feedback]

The sequence below traces a single invoice through the pipeline — this is the path that runs on 90%+ of lines at steady state.

sequenceDiagram participant Carrier participant Nexa participant Contracts participant Atlas participant ERP Carrier->>Nexa: Invoice (EDI/PDF) Nexa->>Nexa: Parse to canonical schema Nexa->>Contracts: Fetch rate card + clauses Nexa->>Atlas: Fetch shipment actuals + POD Nexa->>Nexa: Line-by-line variance check alt within tolerance Nexa->>ERP: Release for payment + remittance else exception Nexa->>Nexa: Package dispute case → Vera end

The decision boundary between auto-release and dispute is configurable per carrier, per service, per variance threshold. Most operators start conservative (tight tolerance) and widen as trust in Nexa’s matching accuracy builds — typically within the first 60 days of steady state.

Why this is different from “freight audit software”

Traditional freight audit-and-pay tools are optimised for batch processing of historical invoices against digitised rate cards. They tell you, 30 days after the invoice lands, that you overpaid. They don’t close the loop on the next invoice, they don’t tie back to the shipment event that actually happened, and they almost never hold the POD in the same system.

Nexa operates on the live event stream. When the shipment is delivered, the ePOD is captured, the GPS trail is recorded, and the dwell timestamps are computed — the settlement-ready data is already in the same system the invoice will land in. By the time the carrier invoice arrives (a day later, a week later, whenever), Nexa doesn’t need to “chase” evidence. She already has it. That shift — from forensic audit to live reconciliation — is why cycle time compresses, why leakage is caught at the line rather than in aggregate, and why the feedback loop to carrier performance actually closes.

Proven outcomes

Customer type & scale	Outcome
Global biotech scaling rare-disease therapeutics across 30+ countries	50–70% manual effort reduction; 95% batch instrumentation; automated invoice reconciliation with rate-card enforcement across multiple carriers
Global alco-bev leader operating across 70+ countries	Fully automated route settlement moved from 70% to over 90%
Southeast Asia’s largest air-logistics network, 164 hubs	Real-time POD sync with back-end; cleaner weekly business reviews on 3PL KPIs
One of India’s largest quick-commerce players, ~1M orders/day	Faster, cleaner settlements; vendor payout automation across own fleet + 3PL partners
One of Asia’s largest quick-commerce arms, 5M+ orders/month	~82% reduction in COD loss; dynamic COD limits, auto-reconciliation, settlement logic

Integrations

ERP & finance: Oracle ERP · SAP · NetSuite · Microsoft Dynamics · Tally (India) · Xero
Treasury & payments: domestic bank APIs · SWIFT · local payout rails (UPI, SEPA, ACH, FAST)
Tax & compliance: GST, VAT, IGST engines · e-invoicing portals (IRN for India, Peppol for EU)
Carrier & 3PL: 240+ carrier APIs · EDI 210/214 · carrier portal scraping for non-integrated vendors
Shipment data: Shipsy TMS · Atlas event bus · ePOD · driver app
Contract & quality systems: Veeva QMS (pharma) · custom contract stores · legal document repositories
Analytics: Snowflake · BigQuery · Redshift · native BI dashboards

Deployment

Nexa goes live in 8–12 weeks for most enterprises; complex multi-country deployments with many carriers extend to 14–16.

Phase 1 — Discovery (weeks 1–2). Inventory of carriers, invoice formats, contract terms, and current leak patterns. Joint workshop with finance, procurement, and logistics.

Phase 2 — Configuration (weeks 2–6). Contract digitisation — rate cards, accessorials, fuel linkage — into the object store. Invoice parser tuning per carrier. ERP integration for payout triggers. Variance tolerance rules.

Phase 3 — Pilot (weeks 6–9). Parallel run with existing audit team for 3–4 weeks. Nexa processes, humans verify, variances are tuned. Success criteria: >95% line-match accuracy, <2% false-flag rate, <5-day cycle time from invoice receipt to payout decision.

Phase 4 — Scale (weeks 9–12+). Rollout to remaining carriers and country entities. Month-end close migration. Governance: weekly variance-tuning review, monthly leakage report, quarterly contract-version audit.

Success criteria locked pre-go-live: settlement cycle time, manual touch rate, leakage-recovered $ value, POD-match completeness. Governance model includes finance ownership of tolerance thresholds and logistics ownership of POD data quality.

Who owns what after go-live

Finance owns the variance policy, approval thresholds, and month-end close cadence.
Procurement / commercial owns the contract object store and renegotiation triggers sourced from systemic flags.
Logistics ops owns POD data quality — the upstream ePOD capture rate, driver-app adoption, and scan coverage that Nexa’s matching depends on.
Shipsy owns engine accuracy, integration uptime, and capability releases.

This four-way ownership is documented in the go-live RACI. When a line fails to auto-settle, the RACI answers “why” and “who fixes it” in the first review.

Security & compliance

SOC 2 Type II · ISO 27001 · GDPR
SOX controls: segregation of duties between matching engine and payout trigger; immutable audit log per decision
21 CFR Part 11 for pharma deployments — electronic signature, system validation
PCI-DSS scope minimisation: Nexa never touches card data; payments flow through ERP/treasury
Full audit trail — every line decision tied to contract version, shipment event, POD, and the rule that fired
Three-tier confidence scoring: auto-release, suggest-release, route-to-human
Regional data residency for contract and invoice data

Case study callouts

Global biotech · rare-disease therapeutics, 30+ countries, ~$4B in net sales

Automated invoice reconciliation, rate-card enforcement, and cost analytics across a fragmented multi-carrier, multi-CDMO landscape. Manual effort reduced 50–70% while SLA compliance climbed past 95%. Integrated with Oracle ERP and Veeva QMS so settlement data feeds batch instrumentation and quality records.

→ Read the full case study

Global alco-bev leader · 70+ countries, 2M+ annual trips

Fully automated route settlement moved from 70% to over 90%. Payments to logistics service providers for excessive stays at customer locations were cut by 28% — a direct consequence of Nexa matching actual dwell time to the contractual grace window, not the carrier’s claim.

→ Read the full case study

One of Asia’s largest quick-commerce arms · 5M+ orders/month, 200+ dark stores

Dynamic COD engine, settlement logic, and auto-reconciliation cut COD loss by ~82%. Cost-per-delivery fell ~21% in parallel as carrier scorecards from Nexa fed back into the allocation engine. Month-end close moved from a finance bottleneck to a weekly cadence.

→ Read the full case study