Freight Audit: The 4 Models, and How to Pick One

Freight audit is checking every carrier invoice against contracted rates and shipment records before you pay it. Enterprises run it through one of four models: manual in-house, audit software, outsourced/BPO, and AI-native full coverage. The model you pick decides how much of the 1.5 to 2.5% of freight spend typically lost to overcharges you actually catch.

‍

Key Takeaways

• Freight audit is a financial control, the formal check that carrier billing matches the rate you signed and the shipment that actually happened, before that money leaves the business.
• Four models exist: in-house manual review, audit software, outsourced/BPO audit, and AI-native full-coverage audit. Each trades off cost, headcount, and coverage differently.
• Most manual and BPO programs sample 15 to 30% of invoices. The other 70 to 85% ships through unchecked, and that is where most recoverable overcharges live.
• A Fortune 500 home appliance manufacturer ran freight audit with a $3,000 invoice threshold that carriers had learned to work around. Moving to full-coverage audit recovered $6M a year, most of it from charges auto-approved for years.
• The right model depends on your freight spend size, your headcount cost, and whether a conflict of interest, like a broker or carrier auditing its own invoices, is already limiting what you catch.

‍

What is freight audit?

Freight audit is checking every carrier invoice against the rate you agreed to, the shipment record, and any accessorial charges, before you pay it. It exists because carrier billing does not fix itself. Rates drift. Accessorials get billed with no proof of service. Nobody on the carrier side is checking in your favor.

The idea itself is simple: does the invoice match what was agreed, and what actually happened. The hard part is doing that check on every single invoice, at real volume, without it turning into a full-time job for a growing team.

‍

Why does freight audit matter?

For a CFO or VP of Logistics, freight audit is not a compliance task sitting outside the P&L. It is the control that decides whether the freight line on that P&L reflects what you actually agreed to pay, or what carriers found room to bill instead. Most manual and BPO programs sample 15 to 30% of invoices, which means the other 70 to 85% of overcharges ship straight through to payment, quarter after quarter, with nobody the wiser.

Full-coverage audit typically recovers 1.5 to 2.5% of total freight spend a year. On $50M in freight, that is $750,000 to $1.25M in cash that was already leaving as an approved overcharge, not a future savings target. It also does something a one-time rate negotiation cannot: it keeps working every invoice cycle, and the discrepancy data it produces strengthens your position in the next carrier negotiation instead of sitting as a number nobody tracked.

‍

What are the benefits and challenges of running freight audit?

The benefits show up once a program has real coverage. The challenges are usually what keeps a program from getting there.

Benefits:

• Faster invoice cycles: full-coverage audit runs in under 3 days, against 30 or more for a manual backlog.
• A stronger negotiating position: discrepancy data by carrier shows exactly where the next contract needs firmer terms.
• An audit trail: every check, exception, and fix logged in one place, which shortens both internal and external reviews.

Challenges:

• Hidden coverage gaps: a team, a vendor, or software can all look like they're working while checking only a fraction of invoices.
• Conflict of interest: a carrier or broker auditing its own invoices, or a vendor paid a flat fee, has little reason to dig deep.
• Documentation gaps: a missing bill of lading or proof of delivery means a charge can't be verified either way, so it often just gets approved.
• Scaling without adding headcount: in-house and most BPO models scale coverage with team size, which caps how much they can check as spend grows.

‍

Where do freight billing errors actually happen?

Freight billing errors concentrate in a handful of predictable places, and almost all of them run in the carrier's favor. Base rate misapplication happens when a carrier bills above the contracted rate for a given lane or mode. Fuel surcharge errors happen when the wrong tier or an outdated index gets applied. Accessorial charges, detention, liftgate, residential delivery, TONU, get billed with no proof the service actually happened, and they carry the highest error rate of any charge category.

Three more patterns round out the list. Weight and dimension errors inflate the billable weight above what actually shipped, which raises the rate before any accessorial even applies. Post-delivery reclassification changes a shipment's freight class after the fact, often based on a re-weigh with no real inspection behind it. And duplicate billing puts the same move on two invoices under different reference numbers, which is easy to miss without cross-invoice matching.

Ocean freight adds a layer most other modes do not carry. Rates get quoted per vendor, per lane, and per time period, often sitting in separate spreadsheets that never sync with each other, and ocean cost is unusually sensitive to fuel price swings on top of that. Most in-house teams handling more than a handful of ocean shipments a month fall back on sampling here specifically, since checking the right rate, for the right vendor, for the right week, by hand, does not scale.

Each of these is a distinct check, not a single pass. For the full line-item breakdown of how to catch each one on every invoice, see how to audit freight invoices.

‍

How does the freight audit process actually work?

The process runs in four stages, regardless of which model you use to run it. First, data collection: pull the carrier invoice together with the contract, the bill of lading, and proof of delivery. Second, matching: check every line item, base rate, weight, class, and accessorial, against those documents and your contracted rate table. Third, exception resolution: flag anything that does not match, then dispute or file a claim on it. Fourth, payment: approve what passed the match, code it to the right GL account, and release it for payment.

Where the four models genuinely differ is in how much of that four-stage cycle happens without a person doing it by hand, and how much of your invoice volume goes through it at all. For the complete walkthrough of setting this process up end to end, see how to run a freight audit process.

‍

Should you audit at the invoice level or the shipment level?

Freight audit can check the invoice as a whole, or the shipment underneath it. The choice matters most for LTL and parcel, where one invoice often bundles many shipments. Invoice-level audit, sometimes called freight invoice validation, checks the invoice against standard weights and package assumptions. It is the simpler approach. It works well when packaging is consistent and volume is high enough that extra complexity is not worth the cost.

Shipment-level audit checks the real weight, dimensions, and measurements behind each shipment. Then it compares that data against the quote and the invoice. It costs more to run, but it catches what invoice-level checking misses, dimensional weight inflation and reclassification especially, since both depend on what the shipment actually was, not what the invoice claims it was. It also ties more cleanly into an ERP, since shipment-level data maps to purchase orders and cost centers instead of one invoice total.

‍

Should freight audit happen before or after payment?

Freight audit can run pre-payment or post-payment, and the difference is what happens to the money in the meantime. Pre-payment audit holds an invoice for review before it clears, so an overcharge never leaves the business in the first place. Post-payment audit lets the invoice pay, then chases a refund or credit afterward if something is wrong.

Pre-payment is the stronger position whenever you can manage it. It stops the cash from moving instead of asking a carrier to send it back, and carriers are, understandably, slower to process a refund than they were to accept the original payment. Post-payment audit still works, and it is sometimes the only option when invoice volume or system speed cannot support a hold-and-review step before every payment run. But it trades a cleaner recovery process for a slower one, and small discrepancies are more likely to get written off rather than chased.

‍

What are the four models of freight audit?

The four models are manual in-house review, standalone audit software, outsourced or BPO audit, and AI-native full-coverage audit. They differ mainly in who does the checking, how much of your volume gets checked, and what it costs.

‍

The real gap between these models is not about technology. It is about how many of your invoices get checked at all. A model that covers 30% of invoices will always miss most of the overcharges sitting in the other 70%, no matter how good it is at catching the ones it does see.

‍

What does the software model actually cost, once IT and manual work are counted?

Rules-based audit software looks automated, but most of what it "catches" still lands back on your team. A static rule can approve a clean match. It cannot resolve an exception, so a variance, a missing reference, or a complex accessorial gets flagged and routed into a queue, not fixed. The work still happens, just downstream of the software instead of before it.

Two costs compound that. Configuring a legacy ERP, SAP or Oracle, to handle multi-modal accessorial schedules, fuel surcharge formulas, or dimensional weight matrices takes custom coding and consultant time, not a settings screen. And without that integration, teams end up doing what amounts to swivel-chair reconciliation: copying reference numbers, PO numbers, and shipment facts by hand across a TMS, an ERP, and a contract system that don't talk to each other.

The software license was never the real cost. The team required to run the queue it generates is.

‍

What's the hidden cost of outsourcing freight audit to a BPO?

A BPO looks like it removes the headcount problem, but it usually just moves the headcount offshore and adds a lag. When you renegotiate a carrier contract, the BPO's staff has to manually update its own rating tables before the audit reflects the new terms. Invoices can clear at the old rate for weeks while that update works through the queue.

The other cost is control. Once the audit is outsourced, your shipment history and carrier contracts live inside the provider's systems, not yours. That usually means no direct visibility into the dispute queue, the audit metrics, or a real-time accrual, just whatever the provider reports back on its own schedule.

{{blue-cta}}

‍

How do you decide between in-house, outsourced, software, and AI-native audit?

You decide based on three things: your freight spend size, the headcount you are willing to dedicate, and whether your current setup has a conflict of interest limiting what it can catch. The table below breaks down what to look at for each.

Freight spend size is usually the biggest driver of the three, since it sets the dollar size of what partial coverage actually costs you. But the conflict-of-interest factor is the one most programs miss, because the audit still looks like it is running.

‍

What does switching models actually recover?

Switching models recovers whatever your current model was structurally unable to catch, which is often larger than expected once you can see it. A Fortune 500 home appliance manufacturer with $115M in annual freight spend ran freight audit with a $3,000 invoice threshold. Invoices below that amount cleared automatically, every cycle, and carriers had learned exactly where that line sat, the kind of systematic pattern an Audit Trends Agent is built to catch going forward.

After moving to full-coverage audit across every invoice, the company recovered $6M annually. The majority came from charges that had been auto-approved for years, not from new errors. The threshold model was not broken. It was doing exactly what it was built to do. It just was never built to catch everything.

‍

What should you evaluate before choosing or switching a freight audit model?

Evaluate coverage first, then speed, then what happens after an error is found. Coverage is the share of invoices actually checked before payment. It is the one number that decides how much of your leak you can ever catch, no matter how good the checking itself is.

• Invoice coverage: what share of invoices get a real line-item check, not a sample or a threshold cutoff.
• First-pass match rate: how often an invoice clears automatically because it matches on the first check, versus needing manual exception handling.
• Speed to resolution: how long it takes from flagging a discrepancy to getting a credit or short-pay applied, not just how fast the flag happens.
• Continuous monitoring: whether an Anomaly Detection Agent catches cost spikes and rate drift as they happen, instead of a month-end report surfacing them weeks later.
• Integration depth: whether the audit model connects directly to your TMS and ERP, or requires manual export and re-entry between systems.

If you are actively comparing vendors against these criteria, the best freight audit software comparison is a faster starting point than evaluating each one from a blank page.

‍

What does mature freight audit coverage actually look like?

Mature freight audit coverage looks different from a working program, but only once you measure it. The table below covers the four metrics that show the difference.

‍

A program can look mature on paper, with a named vendor and a dashboard, and still sit in the immature column on every one of these numbers. The dashboard is not the evidence. The coverage percentage is.

‍

How does freight audit fit into transportation spend management maturity?

Freight audit is the foundation that transportation spend management (TSM) maturity gets built on. It is not a separate initiative running alongside it. TSM maturity is a measure of how far a company has moved away from reactive, siloed freight cost management. A mature company has one view of spend, updated all the time, across every carrier, mode, and region.

A program stuck at 15 to 30% audit coverage cannot get much more mature than that, no matter how good its rate benchmarking or carrier scorecards look. The data feeding those higher-level TSM tools is only as clean as the audit underneath it. Spend visibility, carrier scoring, and rate benchmarking across your network are only as trustworthy as the audit data holding them up. Full-coverage audit is what makes that data trustworthy in the first place.

‍

What's next for freight audit?

The direction of freight audit is toward less human handling of routine work, not more automation bolted onto the same manual process. Rules-based software already moved audit past pure manual review. The next shift is agentic AI that resolves the exception it finds, instead of flagging it and handing the work back to a person.

That shift is not speculative. It is already reflected in how the category is scored. The 2026 Gartner Market Guide for Freight Audit and Payment names autonomous, full-coverage audit as the direction vendors in this space are being measured against, not a future feature on a roadmap.

‍

Freight audit only works if the model behind it can see everything

Most freight audit failures are not audit failures. They are coverage failures wearing an audit process as a disguise. A team checking exceptions by threshold, a BPO vendor working a flat fee, a broker auditing its own invoices: each one is running a real audit process. None of them can see the whole picture, and the money sitting outside what they check does not stop leaking just because nobody is looking at it.

Freehand runs freight audit as a single freight audit and payment software layer with no sampling and no threshold gap. Every invoice gets checked against your contracted rates, your shipment records, and your accessorial rules, and the exceptions get resolved automatically instead of queued for a person. You get the coverage the other three models were never built to deliver, without adding headcount to get there.

When an invoice fails that check, the dispute management agent compiles the packet, contracted rate documentation, shipment proof, and audit trail, and routes it to the right carrier channel automatically. Clean charges get coded straight to your ERP by the GL coding agent, with no manual journal entries behind them. Through Freehand Studio, your team sets tolerance thresholds and approval rules in plain language, with full visibility into the exact contract clause or shipment fact behind every decision, not a black box you have to take on faith.

{{brown-cta}}

‍

Frequently Asked Questions

What is the difference between freight audit and freight payment?

Freight audit checks that an invoice matches contracted rates and shipment records. Freight payment is the actual disbursement of funds to the carrier once that check clears. Many providers offer both, but they are two distinct steps, not one process.

Which freight audit model is most common at large enterprises?

Outsourced/BPO and audit software are the two most common models above $50M in freight spend, mainly because in-house teams rarely scale to that invoice volume. AI-native audit is the fastest-growing model at this spend tier, since it scales coverage with volume instead of headcount.

Can a broker or carrier run freight audit on its own invoices?

Yes, and it happens often, but it creates a structural conflict of interest. A broker or carrier negotiating your rates has little incentive to aggressively flag its own overcharges. This does not make the audit fraudulent. It limits how much it will ever catch.

What percentage of freight invoices typically have billing errors?

Estimates commonly run 10 to 25% of invoices with some billing error. The dollar impact depends heavily on coverage, though. A model that only checks 20% of invoices reports a much lower error rate than one checking 100%, simply because it never sees most of the errors.

Do you need a TMS before you can run freight audit?

No. A TMS manages shipment execution, and it often holds the rate data an audit checks against, so integration helps. But freight audit itself can run against invoice data, contracts, and shipment records on its own, with or without a TMS in place.

How do you know if your current freight audit model has a coverage gap?

Check your invoice coverage percentage directly, not your error catch rate. A high catch rate on a low coverage base still means most errors go unseen. If your model samples, uses a dollar threshold, or has ever been described as auditing "the big ones," a coverage gap almost certainly exists.

What does a freight auditor actually check?

A freight auditor checks the base rate against the contract, the weight and class against the shipment record, and every accessorial against proof it was earned. Duplicate invoices and late reclassifications get checked too, since both change what should have been billed.

Why do carriers make billing errors in the first place?

Most carrier billing errors are not deliberate fraud. They come from manual rate entry, outdated fuel or accessorial tables, and re-weigh processes with limited oversight. What matters for you is that these errors run almost entirely in the carrier's favor, whatever the cause.

Which freight modes should a freight audit program cover?

All of them. LTL, FTL, parcel, and ocean or air freight each have their own billing patterns and error types. A program that covers only one mode, usually LTL, leaves every other mode's overcharges sitting completely unchecked.