New research from Forrester, commissioned by Basware, digs into why AI in accounts payable so often stalls at the pilot stage, even at organizations that did everything right on paper. The short answer: it's not a strategy problem. Most finance teams already have the AI principles, leadership backing, and risk frameworks in place. The gap sits in what happens next, the operating model that must run AI day after day, invoice after invoice, entity after entity. Here's what the research found, why it matters if you lead finance or AP, and where to go for the full picture.
What you need to know
Finance teams have done the upfront thinking. Most have deployed AI somewhere in AP already, usually in fraud detection or reporting, and plan to invest further. Documented AI principles, executive sponsorship, and risk frameworks are common.
What's rare is the structure to run all of it consistently. Far fewer teams have found a working balance between governance and innovation, and even fewer have a dedicated team whose job is to own that balance. Without it, AI in AP ends up as a set of disconnected pilots: one model for fraud, one tool for coding, no one accountable for the system as a whole.
That gap gets sharper for organizations running AP across multiple ERPs and countries. A strategy approved centrally still must hold up entity by entity, invoice by invoice, especially as more regions roll out mandatory e-invoicing and reporting rules.
Why it matters to you
If you lead finance or AP, this changes the question you should be asking your team. Not "how much AI have we deployed," but "who can explain how each AI decision got made and prove it to an auditor if asked."
That question matters more every quarter. Regulators are moving faster than most operating models can keep up with. An AI decision that can't be explained isn't just a process gap anymore. It's a compliance exposure.
It also reframes what "more AI" should mean. The instinct when a program stalls is to add another model or another vendor. The research suggests the opposite works better: teams that treat the operating model itself as the priority are the ones who turn strategy into results they can actually defend.
We call this approach Governed Autonomy: autonomy you can defend, built on a named owner for every AI use case, a standing review of what AI is allowed to decide, and an audit trail that holds up before a regulator ever asks for it. Not a feature. A way of running AI in AP that scales trust instead of assuming it.
Read more
The full picture, including where finance leaders say the gap is widest and what it takes, is in Basware's original article. Read it here: The AI readiness gap that's holding AP back.
