Automating payables and receivables can cut costs sharply, but the software only does what the data says. Three industry voices explain what changed for their finance work and what broke on the way.

Few finance projects promise as much as accounts payable automation. Invoices arrive, get read, matched and routed to an approver without anyone typing them in. Payments go out on time, and receivables match themselves to the bank. Then the rollout starts, and old data problems surface fast.

So we asked industry leaders one question: what has automating accounts payable or receivable changed for your finance team, and what went wrong during the rollout? We welcomed time saved, error rates and cash flow results. Three respondents answered, and each pointed to a failure that the software did not flag on its own.

What the benchmarks say

The gap between strong and weak teams using accounts payable automation is wide. In its 2026 AP benchmark, sponsored by Basware, Ardent Partners found best-in-class teams processed an invoice for $2.67, against $12.50 for everyone else. They approved invoices in 4.1 days rather than 15.4. In addition, 55% of their invoices went through without a person touching them, compared with 31% for the rest.

However, accounts payable automation does not guarantee those results. Even on Ardent’s averages, about one invoice in five still becomes an exception that needs manual work. The respondents below explain where accounts payable automation exceptions come from, and how to prevent them.

The vendor list decides the outcome

Our first respondent, a CFO who now runs a fractional finance practice, described the most common failure in accounts payable automation projects.

“The biggest change I’ve seen as a CFO is that payables automation puts everything through one central point. Invoices are processed automatically and sent to the right approver, so the finance team only deals with the exceptions.

Our problems started because we switched the tool on without a clean vendor list. We had three versions of the same supplier, so invoices went to the wrong person for approval and didn’t match. The software does what the data says.

Before you go live, clean up the vendor list and agree the approval process on paper. Then run the old and new processes side by side for one cycle.”

Ankit Sarawagi, Curator, CFO Matrix

His point matches benchmark data. APQC puts the median share of AP payments that are duplicate or wrong at 1.5%, and even top performers report 0.8%, CFO.com reported. APQC names poor-quality vendor master data among the most common causes. Therefore, cleaning the vendor file is the first step in accounts payable automation, not a nice extra.

When a green tick hides a problem

Our second respondent built a finance tool on his own IT firm’s books, and he described a failure that looked like success.

“I built Yupana on 115 months of my own IT firm’s real figures, so I’ve seen bank automation fail quietly. Our CAMT bank import reported success while it was dropping the counterparty field, because a newer CAMT version puts that field one level deeper in the XML.

A simple count of how many fields the import recognised caught it, and that count now runs with every import. Never trust a green tick on its own.”

Jens Hagel, Co-Founder, Yupana

His example checks out. CAMT is the ISO 20022 format banks use for electronic statements. In version 02, the counterparty name sat under one tag. In version 08, it moved one level deeper. German banks switched customers from version 02 to version 08 in November 2025. A parser built for the old layout can still read amounts and dates, so the import looks fine while names go missing.

That matters for accounts payable automation and any other matching that relies on names, from cash application to bank reconciliation. As a result, a simple field count is a cheap control for both payables and receivables.

Reminders before matching

Our third respondent looked at receivables from a small business angle.

“Linking my invoicing tool to my accounting software automated my receivables, and it saves me about three hours a month. When an invoice is paid, it feeds straight into my profit and loss and my tax estimate, so I no longer match a spreadsheet to the bank statement by hand.

The rollout went wrong for a founder I advised. They switched on automatic payment reminders before part payments were being recorded, so clients who had already paid got chasers and were annoyed. We paused the reminders and fixed the matching first.”

Lilach Bullock, AI Implementation Consultant and Fractional CMO, Lilach Bullock

Her example comes from a much smaller business than our brief targeted, but the lesson scales. Order matters. Matching has to work before reminders go out. This will matter more in the UK, where a late payments bill introduced in May 2026 would cap large firms’ payment terms to small suppliers at 60 days and make interest harder to waive.

Fraud follows the automation

Centralising payments through accounts payable automation also creates a single target. In the AFP’s 2026 fraud survey of US organisations, 74% were hit by business email compromise in 2025. Vendor impersonation is rising, and only 17% of organisations use AI to fight payments fraud.

So any accounts payable automation project needs a hard control on bank-detail changes. A call-back to a known contact, logged in the system, is still the simplest defence. For more on how fraud teams now spot scams, see our roundup on AI fraud models.

How to roll out accounts payable automation

Taken together, the answers suggest a short order of work. First, clean the vendor master and merge duplicates. Next, agree approval rules in writing before configuring them. Then run old and new processes in parallel for one cycle. Finally, add checks that count what each import and match should produce, rather than trusting a success message.

Software vendors are also moving towards fuller automation. Intuit recently brought an automated books close to accounting firms, while Stripe launched free instant B2B transfers between its users. Both make clean data more valuable, not less, for any accounts payable automation project.

What to watch next

E-invoicing mandates will push more firms into accounts payable automation. France began requiring all businesses to receive e-invoices in September 2026. The UK plans mandatory e-invoicing for VAT invoices from April 2029, and the EU’s VAT in the Digital Age rules require e-invoices for cross-border B2B trade from July 2030.

Above all, the respondents agree on one point about accounts payable automation. Automation does not fix bad data. It processes it faster.

FintechBits covers payments, banking and fintech developments for readers in the US and UK. This article is for information only and is not financial advice. Views expressed are those of the FintechBits editorial team.