Logistics Audit & Dispute Intelligence

Carrier billing
errors are systemic.
Prove them.

TRACIUM is a deterministic logistics audit and dispute platform. It ingests carrier invoices and shipment data, produces evidence-backed findings, and executes financial recovery — without dashboards, synthetic analytics, or fabricated conclusions.

100%
Deterministic Findings
5
Integrated Modules
0
Fabricated Analytics
BLACKLIGHT — Audit Engine
TRACIUM BLACKLIGHT — Canonical Audit Engine
AUDIT EXPOSURE IDENTIFIED
$9,575.58
Deterministic Audit Findings
Every finding is a provable discrepancy between SAP intent and carrier claim — not a flag, not a model output.
Backend-First Financial Truth
All financial surfaces derive from canonical backend state. No UI-side inference. No parallel computation.
Recoverable Exposure Control
Disputes are generated from canonical findings and tracked through full lifecycle to confirmed resolution.
Zero Fabrication Architecture
No invented linkages. No synthetic conclusions. If evidence is incomplete, the system reports it — not a finding.
Enterprise Workflow Discipline
Full audit trails. Review before sending. Lifecycle state from OPEN through RESOLVED — nothing deleted, everything traceable.
The Problem

Carrier billing errors are not exceptions.
They are the default.

Carriers operate complex billing systems across service lines, surcharge categories, and weight methods. Your ERP records expected costs at shipment time. These two systems rarely agree — and the difference is almost always absorbed by the shipper.

Across an enterprise processing hundreds of invoices per month, even a 1–3% billing error rate represents material, recurring financial leakage. The problem compounds because most teams have never measured it against a structured audit baseline.

Most tools tell you what you spent.
TRACIUM tells you what was wrong, provable, and recoverable.
No Systematic Invoice Validation
Root Cause

Most teams lack a deterministic comparison between what the ERP expected and what the carrier invoiced. Errors stay invisible because no one has built the infrastructure to find them at scale.

Dispute Processes That Don't Scale
Operational Gap

When errors are found, disputes live in email threads, spreadsheets, and manual follow-up. No audit trail. No status visibility. No structured measurement of what was recovered.

Dashboards Show Spend — Not Errors
Tool Gap

BI tools and spend analytics platforms tell you what was paid. They cannot tell you what should have been paid, or identify the specific invoice row that contains a provable billing error.

Leakage That Compounds Silently
Financial Exposure

Manual review is too time-consuming to run on every invoice. Teams catch the obvious outliers. Duplicates, weight discrepancies, and rate mismatches accumulate — undetected and unrecovered.

How It Works

Ingest. Audit. Dispute. Recover.

TRACIUM is a pipeline, not a tool. Each step produces a structured output that feeds the next — no manual handoff, no spreadsheet intermediary, no interpretation layer between the data and the dispute.

01
Load carrier invoice and shipment data
Upload carrier invoice files in CSV format alongside shipment-side records from SAP or equivalent ERP. TRACIUM ingests, normalizes, and reports on every row — including what linked, what didn't, and why. Unlinked rows are displayed honestly; they produce no findings and generate no disputes.
Upload + Linkage Layer
02
Execute deterministic audit
BLACKLIGHT runs the audit engine against linked shipment-invoice pairs. Overcharge thresholds, duplicate detection, weight comparison, service level validation — each rule fires against evidence. No estimates. No model predictions. Canonical findings only, each with financial value and lifecycle state.
BLACKLIGHT — Audit Engine
03
Investigate findings and surface exposure
Each finding shows the rule that fired, the severity, the full evidence chain, and the discrepancy amount. FINANCIALS surfaces invoice-level financial truth — every row, every variance. SIGNALFORGE provides statistical anomaly context and investigation guidance. It is an explanation layer, not an audit authority.
BLACKLIGHT + FINANCIALS + SIGNALFORGE
04
Generate and review disputes
CLAIMFORGE converts OPEN findings into structured, carrier-facing dispute drafts — pre-populated with finding evidence, raw shipment identifiers, and discrepancy details. Teams review before sending. Nothing is transmitted without explicit approval. Manual, batch, and automated dispatch modes are all supported.
CLAIMFORGE — Recovery Pipeline
05
Track recovery through to resolution
Disputes move through DRAFT → SENT → RESOLVED. Status is tracked continuously and reflects backend truth. Recovery amounts are confirmed outcomes only — never projected. Full finding and dispute lifecycle history is always visible and never deleted.
CLAIMFORGE — Lifecycle Tracking
Platform Modules

Five systems. One audit pipeline.

BLACKLIGHT
Canonical Audit Engine & Investigation Surface
BLACKLIGHT is the system spine. It executes deterministic rules against linked shipment-invoice pairs, produces canonical findings with financial values and lifecycle states, and provides a full investigation interface. Every finding shows the rule, the evidence, the discrepancy amount, and the operational context. It is the sole source of audit truth in TRACIUM — everything else derives from it.
Findings lifecycle: OPEN → DISPUTED → DISMISSED → RESOLVED. Nothing deleted. Every transition tracked.
CLAIMFORGE
Recovery Pipeline & Carrier-Facing Dispute Workflow
CLAIMFORGE converts OPEN audit findings into structured, evidence-grounded carrier disputes. Every dispute derives from a canonical BLACKLIGHT finding. Payloads are pre-populated and reviewed before sending. Manual, batch, and automated dispatch are all supported. Lifecycle state — DRAFT, SENT, RESOLVED — is tracked continuously and drives all financial recovery reporting.
CLAIMFORGE is not a reporting layer. It is a recovery execution system with direct carrier-facing output.
SIGNALFORGE
Statistical Anomaly Layer & Investigation Support
SIGNALFORGE runs statistical anomaly detection independently of the rule engine, surfacing patterns that deterministic rules cannot catch — near-duplicate billing, cost-per-kg spikes, carrier behavior shifts. It is strictly an explanation and investigation guidance layer. It never creates findings, never modifies audit results, and never determines financial truth. This boundary is enforced architecturally.
SIGNALFORGE assists investigation. BLACKLIGHT determines audit truth. These are different systems.
FINANCIALS
Invoice-Level Exposure & Financial Truth Surface
FINANCIALS displays raw invoice-level data exactly as ingested — every row, every variance, every tag. The Exposure Heatmap visualizes audit yield by carrier and rule type. Top discrepancies surface the highest-value individual recovery targets. No aggregation conceals the truth. All figures are backend-driven consequences of canonical audit findings, not computed by the UI.
Finance teams use FINANCIALS to validate disputed amounts against finding evidence and track recovery for period close.
OPERATIONS
Shipment Execution Context & Audit Baseline
OPERATIONS provides the shipment-side baseline that the audit engine compares against carrier invoices. Shipment records carry carrier assignments, service levels, expected costs, route data, package weights, and risk scores. High-risk shipments are flagged based on billed-to-expected divergence. This is the SAP INTENT side of every finding — the ground truth that makes audit possible.
Logistics teams review OPERATIONS to validate that findings reflect real shipment-side discrepancies before authorizing disputes.
Architecture Principle

All five modules share one canonical backend. No parallel computation. No synthetic linkage. Every number on every screen is a direct consequence of deterministic audit truth.

Schedule Platform Walkthrough
Why TRACIUM Is Different

Audit infrastructure.
Not a reporting tool.

TRACIUM is not a BI platform. It does not aggregate spend, model projections, or surface configurable KPIs. It is built at the architectural level to produce evidence-backed findings, enforce audit truth, and execute financial recovery. These are not feature distinctions — they reflect a different design philosophy entirely.

Generic Logistics Analytics & BI Tools
Show carrier spend — cannot identify which specific invoice rows are wrong
Generate flags and alerts with no lifecycle state, no evidence chain, no dispute path
Configurable KPI dashboards — metrics with no underlying audit evidence
No deterministic linkage between invoice rows and shipment records
No dispute generation — every recovery action requires manual human effort
AI-derived conclusions without auditable evidence chains
Recovery measured as projected or modelled savings — not confirmed outcomes
TRACIUM
Deterministic findings — specific invoice rows, specific discrepancies, specific evidence
Canonical findings with lifecycle state: OPEN → DISPUTED → DISMISSED → RESOLVED
All metrics are direct outputs of audit findings — no configured KPIs, no synthetic surfaces
Typed, deterministic identifier linkage between every invoice row and its shipment record
CLAIMFORGE: dispute generation, review, transmission, and tracking — built into the platform
Zero fabrication: if evidence is insufficient, the system states it rather than producing an unsupported finding
Recovery measured from confirmed dispute outcomes only — never modelled or projected forward
Recoverable Conditions

What the audit engine detects.

TRACIUM identifies billing discrepancies where deterministic rules fire against linked evidence. The following are the core recoverable finding categories the audit engine is built to detect and quantify.

Overcharge
Rate & Amount Overcharge
Invoiced amount exceeds the expected cost in the shipper's ERP beyond the defined threshold. Quantified as both a percentage variance and an absolute dollar discrepancy per invoice row.
Duplicate Billing
Duplicate Invoice Detection
The same shipment billed twice under different invoice numbers. Detected through deterministic cross-referencing of shipment identifiers across invoice records. Grouped as one canonical finding.
Weight
Weight Discrepancy
Carrier-billed weight differs from the shipper's recorded gross weight. May originate from re-weigh, dimensional weight calculation differences, or data entry errors. Detected at the invoice row level.
Service Mismatch
Service Level Mismatch
A shipment booked at one service level is invoiced at a higher rate class. The invoice references the correct shipment — the error is in the rate applied, requiring service-level comparison logic to detect.
Responsibility
Billing Responsibility Error
Carrier charges are directed to an account that should not be responsible for that shipment. Without shipment-level billing responsibility visibility, these charges are absorbed rather than challenged.
Rate Application
Discount & Rate Misapplication
Negotiated discounts or contracted rates are not correctly applied to specific invoice rows. The error is structural — in the rate layer — requiring rate-level comparison against expected pricing to surface.
Accessorial
Accessorial Anomalies
Surcharges and accessorial fees appear on invoices without a corresponding shipment-level trigger. Surfaced through SIGNALFORGE's statistical layer as anomalous charge patterns warranting investigation.
Near-Duplicate
Near-Duplicate Billing Patterns
Two invoices with matching billed amounts submitted within a short window for different shipment references. Not confirmed duplicates — statistically unusual patterns surfaced by SIGNALFORGE for investigator review.
Platform Tour

The platform in operation.

Every screen reflects backend state only. What you see is what the audit engine found — no configured views, no synthetic surfaces.

TRACIUM BLACKLIGHT — Canonical Audit Engine
BLACKLIGHT
Canonical Audit Engine
The tri-panel investigation surface — findings list, context intelligence, and investigation view simultaneously. Every finding is severity-ordered, linked to its evidence chain, and in a lifecycle state that drives the entire platform.
Findings sorted by severity with financial exposure per finding
OPEN / DISPUTED / RESOLVED badges reflect live backend lifecycle state
Context panel shows SAP linkage type, root cause, and suggested fix
Investigation view shows the exact invoice–shipment evidence chain
CLAIMFORGE — Recovery Velocity
CLAIMFORGE — Dispute Board
CLAIMFORGE
Recovery Pipeline
Recovery Velocity tracks dispute value across DRAFT, SENT, and RESOLVED over time. The dispute board shows every active dispute with carrier, invoice, service level, and canonical finding link — all at a glance.
Recovery Velocity: live dispute book by created timestamp and lifecycle state
DRAFT / SENT / RESOLVED columns with real-time totals and amounts
Every dispute is pre-populated from finding evidence and reviewed before sending
Canonical link on every dispute traces back to the originating BLACKLIGHT finding
TRACIUM SIGNALFORGE — Signals & Anomalies
SIGNALFORGE
Signals & Anomaly Detection
SIGNALFORGE runs independently of the rule engine, identifying patterns deterministic rules cannot catch. Near-duplicate billing, cost-per-kg spikes, unusual charge concentrations. It provides investigation guidance only — it never creates findings or modifies audit truth.
Priority anomalies with carrier, invoice pair, type, and severity score
Near-duplicate and cost/kg spike detection across all active carriers
Statistical confidence score — HIGH / MEDIUM — with investigation narrative
Runs independently: no OpenAI key required, fallback explanations always provided
FINANCIALS — Exposure Heatmap
FINANCIALS — Invoice Table
FINANCIALS
Invoice Intelligence
Raw invoice-level financial truth — every row, every variance, every OVERCHARGED tag. The Exposure Heatmap shows audit yield sized by carrier and rule type. Top discrepancies surface the highest-priority individual recovery targets.
Exposure Heatmap: carrier-rule exposure from canonical open audit findings
Invoice table with expected vs billed, variance %, and OVERCHARGED tags
Top discrepancies panel identifies highest-priority individual recovery rows
Total overcharged, count, and match/under always visible in sidebar
TRACIUM OPERATIONS — Shipment Intelligence
OPERATIONS
Shipment Intelligence
The shipment-side baseline that grounds every audit finding in operational reality. Each shipment record shows carrier, route, billed vs expected, weight, invoice count, and audit risk score — giving logistics teams the context to validate findings before disputes are submitted.
Shipment list with AUDITED status, carrier, route, and risk score
HIGH RISK / LOW RISK scoring from billed-to-expected divergence
Invoice counts per shipment reveal multi-invoice complexity immediately
Total shipments, high-risk count, average weight, and pending audit visible at a glance
Analytics Philosophy

Every number is
a finding consequence.

TRACIUM does not produce configurable KPIs, modeled projections, or abstract metrics. Every number, chart, and visual on the platform represents something provable, recoverable, and operationally real.

All analytics surfaces are direct visualizations of canonical audit findings, dispute lifecycle state, and recoverable financial exposure. If no findings exist, there is nothing to report. This is not a quality standard — it is a system constraint.

Most tools answer: "What did we spend?" TRACIUM answers: "What is wrong, what is in dispute, and what has been confirmed recovered." The distinction is not cosmetic. One produces reports. The other produces recovery.

METRIC WHAT IT ACTUALLY REPRESENTS
Audit Yield Sum of discrepancy amounts across OPEN findings. A direct output of the audit engine — not a configured KPI.
Recovery Rate Confirmed carrier credits from resolved disputes only. Not projected. Not modelled. Not estimated.
Carrier Leakage Distribution of actual detected errors by carrier and rule type — what the deterministic engine found, not modelled behavior.
Linkage Coverage Percentage of invoice rows linked to shipment records. Low coverage means audit yield is not meaningful yet — fix data first.
Finding Lifecycle Operational pipeline state across OPEN, DISPUTED, DISMISSED, RESOLVED. A direct reflection of the audit and recovery pipeline.
Who It's For

Built for teams with financial accountability for carrier costs.

Logistics & Transportation Teams
Responsible for freight cost control across multiple carriers and service types. TRACIUM replaces ad-hoc manual review with a systematic audit engine that surfaces findings your team can investigate, validate, and escalate with full operational context.
Finance & Accounts Payable
Accountable for invoice accuracy and period-close reporting. Finance teams use FINANCIALS to validate that billed amounts match expected costs at the row level — and to confirm disputed amounts align with canonical finding evidence before authorizing recovery actions.
Carrier Billing & Freight Audit Teams
Dedicated teams responsible for identifying and recovering discrepancies. TRACIUM replaces manual audit with a systematic engine — producing evidence-backed findings on every invoice, generating dispute communications automatically, and tracking recovery through to resolution.
Supply Chain Operations Leadership
Executives accountable for logistics performance and cost control. TRACIUM's audit yield and recovery pipeline data provides the baseline measurement of carrier billing accuracy — and the infrastructure to improve it systematically rather than through periodic manual reviews.
Audit & Compliance Teams
Teams requiring traceable evidence and maintained audit trails. TRACIUM preserves full finding lifecycle history — findings are never deleted, every status transition is tracked, and dispute evidence is exportable for regulatory, contractual, or internal audit purposes.
Enterprise SAP Environments
Organizations running SAP ERP as their shipment system of record. TRACIUM is purpose-built for SAP INTENT vs CARRIER CLAIM comparison. Shipment data loads from SAP exports — no live API integration required to begin generating findings. Future phases include controlled SAP writeback for approved corrections.
Platform Integrity

Evidence first.
No fabrication.
Full traceability.

TRACIUM's credibility depends on one principle: never stating more than the evidence supports. Every finding is backed by evidence. Every financial figure is derived from canonical backend state. If evidence is incomplete, the platform says so explicitly — it does not produce a finding it cannot defend.

These are not aspirations. They are architectural constraints enforced at the system level.

Deterministic Audit Logic
All findings produced by rule-based logic against linked evidence. No probabilistic inference. No AI-determined financial conclusions. Rules fire or they don't.
Full Evidence Traceability
Every finding links to the specific invoice row, shipment record, rule, expected value, and actual value. The evidence chain is always visible — nothing inferred without backing data.
Dispute Review Before Transmission
Disputes are generated as drafts for team review. Nothing is transmitted to carriers without explicit approval. Automated sending is a deliberate configuration choice, not a default.
Raw Identifier Integrity
Shipment, delivery, and invoice identifiers are preserved exactly as loaded. No reformatting, no synthesis. Carrier-facing disputes always use the original identifier values.
Complete Lifecycle History
Findings are never deleted. Every lifecycle transition — OPEN, DISPUTED, DISMISSED, RESOLVED — is preserved for compliance, retrospective review, and audit trail purposes.
Frequently Asked

Common questions, answered directly.

Is TRACIUM a dashboard or BI tool?+
No. TRACIUM is a logistics audit and dispute platform. It does not produce configurable dashboards, abstract KPIs, or modeled analytics. Every data surface is a direct reflection of canonical audit findings, dispute lifecycle state, and recoverable financial exposure. If no findings exist, there is nothing to display. This is a system constraint, not a feature setting.
How are audit findings generated?+
Findings are produced by BLACKLIGHT, TRACIUM's deterministic audit engine. The engine applies defined rules — overcharge thresholds, duplicate detection, weight comparison, service level validation — against linked shipment-invoice pairs. A finding exists because specific field values from two sources differ in a way that satisfies a defined rule. There is no probabilistic inference and no AI-determined conclusion.
What data is required to get started?+
Two data sources are required: carrier invoice files in CSV format (invoice number, shipment reference identifiers, billed amounts, service level fields), and shipment-side records from SAP or equivalent ERP (shipment number, delivery number, expected cost, carrier, service level, weight). A client mapping layer is configured during onboarding to align your field names with TRACIUM's canonical model.
Does TRACIUM use AI?+
TRACIUM includes SIGNALFORGE — an AI layer used for explanation, statistical anomaly detection, and investigation guidance. SIGNALFORGE is not the audit engine and is not a source of findings. Audit findings are generated by BLACKLIGHT using deterministic rules, not AI inference. SIGNALFORGE interprets what the deterministic engine found. It cannot create, modify, or override findings. This boundary is enforced architecturally.
Are disputes reviewed before going to carriers?+
Yes. CLAIMFORGE generates dispute drafts pre-populated with finding evidence — reviewed by your team before any transmission. Manual, batch, and automated dispatch are all supported. Automated sending is a deliberate configuration choice, not a default. Nothing is transmitted to a carrier without explicit user action.
How does this compare to using a third-party audit firm?+
Third-party audit firms review invoices in arrears, charge 25–50% of recovered amounts, and deliver findings as reports. TRACIUM audits continuously, produces findings as structured data with lifecycle state, generates dispute communications within the platform, and costs a fraction of contingency fees at scale. Your team controls the findings and acts on them directly.
What happens to invoices that don't link to shipments?+
Unlinked invoice rows are displayed in FINANCIALS with their linkage status and the reason the link failed. They produce no audit findings and generate no disputes. TRACIUM does not fabricate findings from unlinked data. Unlinked rows may indicate data quality issues, format mismatches, or genuinely unrecognized references requiring investigation before audit can proceed.
What does implementation look like?+
Implementation begins with a client mapping layer that maps your field names to TRACIUM's canonical model — typically completed in 2–4 weeks using representative invoice and shipment samples. No live ERP integration is required to begin. Initial operation is CSV-based. First audit findings are typically available within 3–4 weeks of onboarding start, depending on data quality and identifier consistency.
Get Started

Audit your carrier billing.
Recover what is owed.

Request a platform demo or book an intro call. We will walk through your invoice data structure, run TRACIUM against a sample, and confirm whether there is recoverable exposure before any commitment.

Direct contact: thomas.rvachew@tracium.tech