How to Handle Non-PO Invoices Efficiently at Scale

Every finance team has a version of the same problem. PO-backed invoices flow through automation cleanly, because the purchase order already carries the approval, the coding and the commitment. Non-PO invoices arrive with none of that. Someone has to work out what the spend was, which cost centre owns it, whether it was authorised, and who is allowed to approve it — and they usually have to do it by email.

At low volume that is an irritation. At scale it is the single biggest constraint on AP productivity, because non-PO invoices consume disproportionate effort per invoice and are the category most exposed to duplicate payment and supplier fraud.

What is a non-PO invoice?

A non-PO invoice is a supplier invoice that arrives without a matching purchase order. Because no PO exists, it cannot be validated by three-way matching against a PO and a goods receipt. Instead, it has to be coded, validated and approved by a person with budget authority — which is why the control model, not the software, determines whether it scales.

Typical non-PO categories include:

• • Utilities, rent, rates and insurance
  • SaaS subscriptions and software renewals
  • Professional services — legal, audit, consultancy, recruitment
  • Freight, logistics and courier charges
  • Emergency repairs and unplanned maintenance
  • Low-value ad-hoc purchases below the PO threshold

What percentage of invoices should be non-PO?

There is no single benchmark, and the published ranges differ by region and industry. Analysis published on SAP Community puts typical non-PO usage at 5–30% of invoice volume and 5–25% of AP spend, and treats anything above 30% as a signal of slow PO processes, heavy services spend or immature procurement systems. IOFM benchmarks cited across the industry put the figure higher, at roughly 30–50% of invoices received — a gap largely explained by stronger PO discipline in North American manufacturing versus services-heavy and European operations.

The practical point is not the benchmark. It is that non-PO share is a KPI in its own right, and a rising share is a procurement warning sign rather than an AP one. The thresholds below are a workable starting position:

Threshold framework adapted from SAP Community analysis of non-PO invoice practice.

Why non-PO invoices break AP at scale

Three things compound as volume grows.

1. Effort per invoice is structurally higher

A PO invoice validates itself. A non-PO invoice needs manual coding, a routing decision and a human approval before it can post. According to Ardent Partners’ Accounts Payable Metrics That Matter in 2025, the average organisation spends $9.40 to process a single invoice and takes 9.2 days to do it, while best-in-class teams reach $2.78 and 3.1 days. Non-PO volume is where most of that gap lives.

2. Exceptions and approval delays are the top reported constraints

In the same research, high exception rates and lengthy invoice approval workflows rank among the most-cited challenges facing AP leaders, alongside rising fraud risk. Best-in-class teams run a 9% exception rate against an industry average of 22%. Non-PO invoices are over-represented in that exception pile because there is no PO to resolve the ambiguity.

3. The fraud surface is wider

Invoice fraud and payment-diversion attempts target exactly the invoices that no automated match will catch. Without a PO, the only barrier between a fraudulent invoice and a payment run is the control set the finance team designs.

How to handle non-PO invoices efficiently at scale: a seven-step framework

Step 1 — Close the side doors

Non-PO invoices that arrive in individual inboxes never enter the process at all until someone forwards them. Route every invoice — PDF, scan, email attachment, XML or EDI — to a single intake point and publish it to suppliers as the only accepted channel. A centralised invoice hub removes the shadow queue sitting in personal mailboxes, which is usually where month-end surprises come from.

Step 2 — Classify at intake, not at approval

Decide what kind of invoice you are holding on arrival, not three days later: PO-backed, contract-backed, recurring non-PO, or ad-hoc non-PO. Recurring non-PO invoices — the same utility, the same landlord, the same SaaS renewal each month — should follow a predefined workflow with no human decision at all. Splitting recurring from ad-hoc at intake typically removes the largest single block of manual work.

Step 3 — Code before you route

GL account, cost centre and tax treatment are the most time-consuming part of non-PO processing, and sending an uncoded invoice to a budget owner guarantees a round trip. Invoice coding automation that learns from historical treatment of the same supplier can apply the coding before routing, so the approver is confirming a decision rather than making one. Reserve AP review for low-confidence suggestions only.

Step 4 — Replace three-way matching with a defined control set

This is the step most teams skip. If you remove PO matching, you have to put something in its place — deliberately, and in writing. The replacement set should cover:

Run all of these before the invoice reaches a budget owner. An approver who is also the first control is not a control.

Step 5 — Route on rules, not relationships

Linear approval chains stall the moment one person is on leave. Automated invoice approvals should route on cost centre, department, entity, supplier and value — with defined thresholds, delegated authority, automatic reminders and escalation when an SLA is breached. Every action timestamped, so the audit trail is a by-product rather than a project.

Illustrative thresholds — set yours against your own delegated authority matrix.

Step 6 — Work exceptions only

The operating principle at scale is that people touch invoices that fail a rule, and nothing else. Unknown supplier, suspected duplicate, budget overrun, low-confidence extraction, missing documentation, changed bank details, high value — those go to a human. Everything else posts. This is the difference between an AP team whose workload grows with invoice volume and one whose workload grows with exception volume.

Step 7 — Shrink the pile itself

Automation makes non-PO invoices cheaper to process. It does not make them a good idea. Audit non-PO volume quarterly, rank suppliers by frequency, and move the recurring ones onto blanket POs, contracts or catalogues. Lightweight PO creation inside the AP platform matters here, because the usual reason teams don’t raise POs is that ERP licensing and complexity make it painful for occasional requesters.

PO vs non-PO invoices: what actually changes

Which KPIs prove the process is working

Measure non-PO invoices separately from PO invoices. Blending them hides the problem, because a healthy PO touchless rate will mask a poor non-PO one. Cross-industry AP benchmarks published by APQC are a useful external reference point once you have a baseline.

If you want to size the opportunity before building a business case, the invoice processing savings calculator gives you a starting figure, and there is more detail on structuring the metrics in our guide to AP performance management.

A 90-day sequence

Four mistakes that stop this scaling

1. Automating capture but not coding. Extraction without coding just moves the bottleneck one step later, onto the approver.
2. Treating approval as the control. A budget owner confirms the spend was theirs. They are not checking for duplicates or bank detail changes — the system has to.
3. No published policy. If nobody has written down which categories may bypass a PO, every invoice is arguable and none can be enforced.
4. Leaving the ERP out of scope. Approved invoices that still need rekeying have not been automated. ERP integration closes the loop.

How Kefron AP handles non-PO invoices

Kefron AP is managed accounts payable automation, which matters most in exactly this category. Non-PO invoices are the ones that generate low-confidence extractions and edge cases, and Kefron’s data assurance team reviews flagged fields before the invoice moves — so the customer’s own AP team is not the fallback when the model is unsure.

Next step

See how Kefron AP codes, validates and routes non-PO invoices without adding headcount — book a demo.

Frequently asked questions

What is a non-PO invoice?

A non-PO invoice is a supplier invoice received without a matching purchase order. Because there is no PO or goods receipt to match against, it cannot be validated by three-way matching and must instead be coded, validated and approved by someone with budget authority.

What is the difference between a PO and a non-PO invoice?

A PO invoice is approved before the spend happens, at the point the purchase order is raised, and carries its coding and authorisation with it. A non-PO invoice is approved after the spend has already been committed, so approval, coding and control all happen at invoice stage.

How do you process a non-PO invoice?

Capture it through a single intake channel, classify it as recurring or ad-hoc, apply GL and cost centre coding before routing, run vendor validation and duplicate and fraud checks, route it to the budget owner based on cost centre and value, then post it to the ERP with a full audit trail.

Can non-PO invoices be automated?

Yes. Capture, coding, validation, routing and ERP posting can all be automated. What cannot be removed is the approval itself, because a budget owner confirming the spend is the control that replaces three-way matching. The goal is to make that approval a single confirmation rather than an investigation.

What percentage of invoices should be non-PO?

Published ranges vary from 5–30% of invoice volume (SAP Community analysis) to 30–50% (IOFM benchmarks), depending on region and industry. A practical target is under 15% of volume and under 10% of AP spend, with anything above 30% treated as a procurement issue requiring remediation.

What controls replace three-way matching for non-PO invoices?

Vendor master validation, duplicate detection across multiple fields, bank detail change screening, anomaly checks on amount and frequency, evidence of receipt such as a contract or service entry, and segregation of duties between coding and approval. All of these should run before the invoice reaches an approver.

How do you reduce non-PO invoice volume?

Audit non-PO volume quarterly and rank suppliers by frequency. Move recurring suppliers onto blanket POs, contracts or catalogues, publish a short list of categories that may legitimately bypass a PO, and make raising a PO easy enough that requesters do not route around it.