How does an AP manager build the case for an audit?

Executive Summary

An AP manager builds the case for a spend audit the same way any internal proposal gets approved: with evidence a skeptical reader cannot wave away, a scope narrow enough to be credible, and an ask framed in terms the sponsor already tracks. The mechanism causing the gap is structural, not a performance problem. Rate cards, volume tiers, and surcharge schedules live in contract PDFs outside the ERP, while AP pays against the PO and the invoice total.

Nothing in that workflow checks the invoice against the contract clause.

What changes the outcome is documentation before advocacy. An AP manager who can point to three specific invoices with the contract clause each one violates has a stronger case than one who says spend "feels high." The documentation does not need to be exhaustive. It needs to be verifiable by someone else in five minutes.

The ask itself should be scoped and time-boxed, not open-ended. A CFO who hears "audit everything" hears a headcount request. A CFO who hears "a fixed-scope review of freight and contract labor invoices against their contracts, done in weeks" hears a decision they can make today.

1. What evidence actually moves a CFO to approve an audit?

Specific, verifiable examples move a CFO. Three invoices where the surcharge exceeds the rate card, with the clause number and the dollar delta attached, are worth more than a spreadsheet of aggregate spend. The evidence has to answer one question before anyone asks it: could someone else check this in five minutes and reach the same conclusion?

If yes, it survives the first round of skepticism. If it requires trusting your judgment, it does not.

Start with a small sample, not a full review. Pull ten to fifteen recent invoices from your highest-spend service vendors and lay each one next to its contract. You are looking for three things: a rate that does not match the rate card, a surcharge that should have expired but did not, and a volume tier that should have triggered a discount and did not.

Document each finding the same way: invoice number, contract clause, what the contract says, what the invoice charged, and the difference in dollars. This format is what makes the case portable. A controller or CFO can hand it to someone else and get the same answer.

Resist the urge to extrapolate from three findings to a company-wide number. That estimate will be the first thing a skeptical reader challenges, and it does not hold up. Present the sample as exactly what it is: a sample that shows the pattern exists and is worth investigating at scale.

2. Which vendor categories should the first sample cover?

Pick categories where the contract terms are complex enough to drift and the spend is large enough to matter: freight and 3PL, contract labor, and maintenance or MRO agreements are common starting points because their contracts carry rate tables, tier triggers, and surcharge clauses that a simple three-way match does not test. Choose two or three categories, not all of them. A narrow, well-documented sample is more persuasive than a shallow scan across every vendor you have.

Limiting the sample to a handful of categories keeps the review manageable inside a normal workweek and keeps the evidence tight enough to defend line by line.

Each category chosen should be one where you or someone on your team already has a working knowledge of the contract structure, since reading an unfamiliar contract type correctly under time pressure is where sampling errors happen.

• Freight and 3PL: Fuel surcharges, accessorial fees, and lane rates are set contractually and change on a schedule that billing systems do not always follow.
• Contract labor and staffing: Bill rates, overtime multipliers, and not-to-exceed caps sit in the staffing agreement, not in the timesheet the invoice is built from.
• Maintenance and MRO agreements: Service contracts often include volume commitments or tiered pricing that only shows up when someone checks the invoice against the agreement directly.

3. How should an AP manager frame the request to the CFO?

Frame the request as a scoped, time-boxed review, not an open-ended audit. Name the categories, the sample size already reviewed, and the specific dollar findings from that sample. State what a fuller review would cover and how long it would take.

A CFO approves requests with a defined start, end, and deliverable far more readily than a request to look into spend, because the first is a decision and the second is a commitment to an unknown.

Lead with the finding, not the request. Stating the specific invoices and the documented dollar gap gives the CFO something concrete before they hear what you want.

Then name the scope you are asking to expand into: which categories, roughly how many vendors, and what the deliverable looks like. Do not ask for a standing audit function. Ask for a bounded review of the categories where your sample already found something.

Finally, connect the ask to something the CFO already tracks. If your company is closing on gross margin, tie the request to the margin bridge. If the sponsor is focused on EBITDA, tie it to the levers already on their list. The request should land as a continuation of a conversation they are already having, not a new one.

4. What should the AP manager not claim in the case?

Do not claim a company-wide dollar figure, a projected recovery total, or a percentage of spend at risk from a small sample. Those numbers invite the challenge of how you know, and the honest answer is that you do not, yet. The credible version of the case states the sample size, the specific findings within it, and recommends a scoped review to determine whether the pattern holds at scale.

That is a claim you can defend in the room.

A sample of ten invoices cannot support a statement about a company-wide percentage of spend at risk. It can support a statement that a documented number of invoices sampled showed a discrepancy against their contract terms. The second version is defensible under questioning. The first is not, and a single unanswerable question in the room can cost you the whole case.

The same discipline applies to timelines and cost. Do not estimate how long a full audit would take or what it would cost unless you have a sourced basis for that estimate. If you are proposing to bring in outside help, let that estimate come from the party doing the work.

The goal of the case is to get the door opened, not to close the deal on the first page.

5. Who else needs to be in the room before the CFO is?

Procurement or the vendor relationship owner should see the findings before the CFO does, not after. They will either confirm the contract read is correct or catch a reading you got wrong, and either outcome strengthens the case. A finding the CFO hears about for the first time in the same meeting where procurement disputes it undermines the whole proposal.

A finding both AP and procurement have already agreed on survives the meeting intact.

This step is easy to skip because it feels like it slows the case down. It does the opposite. A CFO who confirms with procurement ahead of time moves faster than one who hears a dispute in real time.

If procurement disagrees with your contract read, that disagreement is worth surfacing before the CFO meeting, not during it. Either you learn something that changes the finding, or you get a documented second opinion that makes the case stronger.

Controller or accounting leadership is the other useful pre-read, particularly if the findings touch categories that affect month-end close or existing accruals. A finding that reconciles cleanly with close activity is easier to act on than one that raises a separate question about the books.

6. Should the AP manager propose doing the full audit in-house?

That decision depends on what the sample already showed and how much AP capacity exists to do the work without disrupting normal payment cycles. An in-house review works when the finding is narrow and the AP team already understands the contract terms involved. A broader review across many vendor categories, done alongside regular AP duties, competes for the same hours the sample took to build, and that tradeoff should be named explicitly in the case rather than assumed away.

Name the tradeoff rather than picking a side for the reader. Doing the full review in-house keeps the work close to the people who know the vendor relationships, but it adds to a workload that is already running the day-to-day payment cycle. An outside review adds a step of onboarding but does not compete with close deadlines or vendor calls.

Either way, the case should state which option you are recommending and why, tied to what the sample found. If the sample surfaced a pattern across many categories, propose the broader scope explicitly rather than letting the CFO infer it from the size of the finding.

This is also where a fixed-scope engagement differs from a contingency-fee recovery firm: the client retains 100% of recoveries, across ValueXPA diagnostics, and the fee does not scale with the recovery, which is a distinction worth naming if the CFO has been pitched contingency models before.

For the wider pattern this sits inside, start with the margin drift guide.

7. Frequently Asked Questions (People Also Ask)

How many invoices should I sample before I raise this with my CFO?

Ten to fifteen invoices across your highest-spend service vendors is enough to show whether a pattern exists. The goal of the sample is not statistical coverage, it is a small, verifiable set of documented discrepancies that someone else can check against the contract in a few minutes.

What if I can't find the actual signed contract for a vendor?

That is itself a finding worth including. An invoice being paid against a rate nobody can currently verify against a signed contract is a control gap, separate from whether the rate itself is wrong. Note it as a distinct item rather than skipping the vendor.

Should I involve procurement before or after I talk to the CFO?

Before. A finding procurement has already reviewed and confirmed is far harder to dispute in the CFO meeting than one they are hearing for the first time. If procurement disagrees with your read, resolve that first.

Can I estimate the total dollar impact across all our vendors from a small sample?

No. Extrapolating a dollar figure from ten invoices to your full vendor base is the first thing a skeptical reader will challenge, and it will not hold up. State the sample findings as exactly what they are and recommend a scoped review to determine the fuller number.

What categories of spend are worth sampling first?

Categories where contract terms are complex enough to drift and spend is large enough to matter: freight and 3PL, contract labor, and maintenance or MRO agreements are common starting points because their contracts carry rate tables and surcharge clauses that routine invoice matching does not test.

Does three-way matching already catch this kind of error?

Three-way matching checks the invoice against the purchase order and the receipt. It does not test whether a surcharge should have expired, whether a volume tier should have triggered a discount, or whether a rate still matches the contract's rate card. Those checks require reading the contract itself.

What's the difference between this and a contingency-fee recovery audit?

A contingency-fee firm typically charges 25% to 50% of recoveries, across ValueXPA diagnostics' market observation. A fixed-scope review is priced upfront and the client retains 100% of recoveries, which changes the incentive on scope and timeline.

How long does a scoped audit typically take once approved?

A prioritized recovery and prevention roadmap can be delivered in 2 to 4 weeks, across ValueXPA diagnostics, once the vendor categories and scope are agreed. That is separate from the sample-gathering stage described here, which is something an AP manager can do internally first.

What should I do if the CFO asks for a dollar estimate I don't have?

Say plainly that the sample was not designed to produce a company-wide estimate and propose the scoped review as the step that would produce one. Offering a fabricated number to satisfy the question costs more credibility than declining to answer it.