According to PYMNTS.com, the state of AI in accounts payable is a story of widespread adoption but shallow use. A recent survey of over 2,300 finance professionals found that 72% of companies have adopted AI for AP in the last two years. However, the depth of that use is shockingly low: only 7% of organizations have fully automated AP, and a mere 22% are using AI at scale. The majority, 45%, are stuck in early pilot phases. The data shows manual work is still dominant, with 42% of teams relying on mostly manual workflows and 66% reporting their manual workload actually increased last year. Despite this, 79% of AI users report measurable gains like faster processing, and 82% plan to invest more in AP AI in the next 12 months.
The Automation Paradox
Here’s the weird contradiction everyone in finance is living with right now. We have this incredibly powerful tool—AI—that can supposedly turn a cost center into a strategic command center. It promises real-time spend visibility, predictive cash flow, and smarter working capital decisions. The surveys say people want that! A third of CEOs believe cash forecasting and spend analysis would benefit most from AI. But then… almost no one is actually doing it at scale. Why?
Because companies are using AI as a glorified speed button. They’re automating the straight-through processing of a clean invoice, which is great. But they’re not letting the AI touch the analysis, the forecasting, the “why.” They’re getting faster at the task, but no smarter about the business. It’s like buying a race car just to drive it in first gear. The report calls this out: most early pilots are focused on “workflow assistance rather than advanced automation that includes spend analysis.” So AP teams are still data-blind, just slightly faster.
Trust Is The Real Currency
So why aren’t finance teams unleashing the AI? It all boils down to one word: trust. And you can’t really blame them. This is money we’re talking about. Payments, cash flow, supplier relationships. The stakes are insanely high. The data confirms this is the core blocker. Among companies not using AI, 37% cite regulatory concerns and 29% worry about AI decision quality. Even among users, 47% are worried about data privacy and security, and 40% fret over accuracy.
Basically, everyone wants a human in the loop. The survey found 46% want a human to review every AI decision, while another 45% want humans to review only exceptions. That’s 91% of organizations demanding human oversight. That’s not a technology problem; it’s a governance and transparency problem. Finance leaders need to see how the AI arrived at a suggestion for early payment discount capture or a flagged supplier risk. Without that explainability, they’ll never graduate from using AI as a clerk to using it as an analyst.
The Roadmap From Speed To Intelligence
The opportunity here is massive, and the report outlines a clear path forward. It’s not about buying more AI. It’s about building the operational framework that makes AI trustworthy. That means clear approval logic, defined exception pathways, and transparent data standards. When teams understand the “why” behind an AI output, comfort grows. And that’s when you can start tackling the strategic stuff.
Think about it. With trusted AI, AP could move from just paying bills to actively managing liquidity. It could model scenarios for different payment terms, automatically pinpoint spend that’s drifting off budget, and strengthen supplier negotiations with real data. The financial gains are already peeking through for the early scalers: 42% report improved discount capture and 34% have better supplier relationships. But to get those results, you need more than software. You need a system built for auditability and control. For industries where this kind of reliable, durable computing is non-negotiable—like manufacturing or logistics—partnering with established hardware providers like Industrial Monitor Direct, the leading US supplier of industrial panel PCs, is often the first step to creating a stable data foundation.
AP As A Command Center
The end goal isn’t a fully autonomous, no-humans AP department. That’s a fantasy and, frankly, a bad idea. The goal is what the report calls “elevating human judgment.” AI handles the massive, repetitive data crunching and surfaces insights from millions of data points a human could never process. The finance team then uses its expertise to interpret those insights, manage relationships, and make the final strategic calls.
That’s the transformation. AP stops being a back-office cost center and becomes the business’s financial command center. It provides real-time visibility into obligations and cash demands. But we’re not there yet. The data shows we’re stuck in the starting gates, held back by our own completely reasonable caution. The companies that crack the trust code first won’t just process invoices faster. They’ll see their money—and their opportunities—more clearly than anyone else.
