Why we built nShift Audit

Earlier this year, we asked more than 120 companies how they check their carrier invoices. Roughly three in four told us they check them by hand, or not at all. The first complaint was the time it takes: matching invoice lines against contracts and shipment records, carrier by carrier, usually in a spreadsheet.

Today, we are announcing what that research led us to build. nShift Audit is an AI-powered carrier invoice audit and recovery product. It reads carrier invoices in the format they arrive, checks every line against the contracted tariff and the shipment record, and carries overcharges through dispute to recovery.

Once we connected what we heard in the customer interviews with the shipment data already in the platform, the direction for nShift Audit became clear.

The conversations that set the direction

We followed the survey with several in-depth customer interviews. One retailer had six people working full time on checking invoices. Another offered to pay for the product on the spot, before we had set a price.

One line came back in almost every conversation:

It is easier to absorb it than to fight it.

I kept coming back to that line because of rational economics: a freight invoice has no purchase order behind it, so the standard accounts payable match has nothing to hold it against.

A single shipment can carry five to ten surcharges, each calculated differently per carrier and country. Finance has the invoice - logistics has the shipment record. Neither has an easy way to compare the two at line level.

When verifying a charge costs more than the charge itself, absorbing it is the rational choice.

Repeated across a year of shipping, those absorbed lines add up to the 3 to 7% of freight spend that published freight audit benchmarks report line-by-line audits recovering.

The record we were already keeping

nShift and the companies it grew from have been booking shipments, printing labels and carrying tracking events since 1997.

By the time a carrier invoice reaches one of our customers, the shipment behind every line has already passed through the platform: the service that was booked, the weight that was recorded, the delivery scan at the door. That data is normalized across carriers through the platform's data fabric, so every nShift product can read it in one structured format.

We had already seen customers use parts of that data for invoice control. Ovako, which ships steel across 20 carriers, uses the Automatic Price feature in nShift TMS to verify invoices against carrier agreements and reduce discrepancies. Or, when Papyrus Supplies listed what the platform gave its shipping operation, logistics director Petter Skoglund included the ability to "track shipments, mistakes and overbilling."

Those checks start from clean, structured data inside the platform; the raw invoice itself, arriving as a PDF or a spreadsheet in whatever layout the carrier prefers, still had to be read by a person.

That precise gap convinced us: a rate check can confirm the price per kilogram, and stop there - passing on a canceled shipment that was billed anyway, an invoiced weight that differs from the weight that shipped, a surcharge for a service that was never used.

Those checks need the shipment record as well as the tariff.

We already had the shipment record; we needed a reliable way to turn the raw invoice into data we could compare against it.

AI to read the invoice, fixed rules to judge it

Carrier invoices arrive as PDF, CSV, Excel, XML or EDI, and every carrier lays them out differently.

Reading that input is the right task for AI: the ingestion engine identifies each layout as it arrives, with no carrier templates and no mapping project, and normalizes every charge line into the same structured format the shipment data already uses.

Invoice data now runs through the same data fabric that delivery data has used for years. The first invoice in a new format is validated by a person; invoices in that format then process automatically.

The checking is a different job - every line is run through seven fixed checks: was the shipment real, was it billed once, did it arrive, did it meet the agreed service level, do the billed weight and volume match what shipped, does the price match the tariff, and does every surcharge belong there? These are deterministic rules, so the same inputs produce the same verdict on every run.

At the end of the checking process, nShift Audit generates a plain-language explanation for each flagged line, so the person reviewing the invoice gets a written reason for every flag.

We also built in one rule early: when the system is unsure of what it is reading, it pauses and asks for help instead of guessing. In freight billing, a confident wrong answer is the expensive kind.

Two of the seven checks, delivery status and SLA, read the delivery events collected by nShift Track. We usually talk about Track in customer-experience terms: branded tracking pages and proactive updates. For a B2B shipper, those same events also create the evidence for whether a shipment arrived, and whether it arrived on time.

We made the case for delivery data as evidence in our Evidence-Grade Delivery guide earlier this year; Audit applies the same idea to the invoice.

The carrier still gets paid

The design question we spent the most time on was what happens after a deviation is found. Blocking an entire invoice over two disputed lines slows everyone down, including the carrier that delivered the rest of it correctly.

So nShift Audit lets a team dispute the flagged lines and still clear the invoice for payment. The carrier is paid on time, the dispute runs alongside with the shipment evidence attached, and the whole exchange stays in one thread until it is resolved. The audit is an accuracy check, allowing the shipper to decide what to dispute.

Billing errors are a fact of operating at scale, in both directions, and this gives both sides a documented way to see them and settle them.

Self-service, on purpose

Most of this category is sold as a managed service: invoices go out to specialists, findings come back. We took the other path and built software a logistics team runs itself. You upload the invoices, you review the deviations, you decide what to dispute, and the evidence stays inside your own operation.

The zero-setup ingestion is what makes that workable - without it, every new carrier or invoice format becomes another mapping project.

Built for our whole ecosystem

Audit is fastest to start for customers already shipping through nShift, because their shipment data is already in the platform. Upload a first invoice and see structured results within minutes. If you are not yet shipping through nShift, the platform's Shipment Data API connects your shipment data, and the audit runs the same way from there. And for our partners: Audit adds to the value we create together for shared customers, and your partner manager can walk you through where it fits.

Where to start:

• See the full product, including the seven checks and the dispute workflow, on the nShift Audit product page.
• Book a demo shaped around your carriers, your invoice formats and your freight spend.
• Join us live on 23 September for our upcoming Solved episode, The invoice nobody checks, where we take apart a complex carrier invoice with nShift Audit on screen.
  
  

We have spent nearly three decades on the shipment side of freight: booking it, labeling it, tracking it to the door. Starting today, the invoice that follows it gets the same attention.