Written by Lyndsey Williams

Interested in a smoother purchasing & AP process?

We help our customers save time, money, and gain control and visibility over purchasing.

If your organization suffers from a disjointed purchasing process, our integrations and workflows can help you get back on track!

Find out more

What a New Prime Minister Means for Business Budget Planning This Autumn

Let's Keep In Touch

Want us to keep you in the loop with all things Compleat?

Ready to know more?

If you’ve read all you need to about how purchasing & AP automation can help you, it’s time for the next step!

Book a Demo

97% of Finance Teams Have Adopted AI. Most Still Can’t Prove It’s Working.

Aug 27, 2026

Finance professional reviewing spend management software and AI tools

Ask any finance leader whether their team uses AI, and the answer is almost always yes. Ask them whether it is actually working, and the confidence drops fast.

According to recent research, 97% of finance departments have adopted AI in some form, up sharply from 76% just a year ago. But only 23% of organisations say AI is exceeding expectations, and a separate poll found 45% of finance teams are still stuck in limited pilot mode, with just 17% actively using it in their core workflows.

That gap between adoption and impact is not a technology problem. It is usually a process problem wearing an AI-shaped label.

Where AI Is Actually Delivering Value

The use cases seeing real traction are specific, not sweeping. Process automation leads at 79%, followed by data visualisation and knowledge management. Within finance teams specifically, the clearest wins are in accounts payable automation, cited by 37% of teams, and error and anomaly detection, at 34%.

That pattern is telling. The AI succeeding in finance right now is not the ambitious, all-purpose kind. It is narrow, applied to a specific, repetitive task, like matching invoices or flagging anomalies, where the process around it was already reasonably well defined.

Why So Many Pilots Stall

68% of CFOs say they have been slow to adopt AI simply because they do not know where to start, and it is easy to see why. Bolting AI onto a finance admin process that is still largely manual, built on spreadsheets, email approvals, and disconnected systems, rarely produces a clean result. The AI ends up automating the mess rather than removing it.

The finance teams seeing measurable impact, only 28% of those with AI investments according to the same research, tend to be the ones who fixed the underlying process first, then applied automation on top of a system that already had structure.

AI works best on top of a process that already makes sense. Spend management software brings structure to purchasing and approvals first, so automation actually has something solid to build on.

Explore Purchase-to-Pay

Starting Where the Data Already Is

Finance teams do not need to solve every use case at once. The teams making progress are starting with the processes that already generate consistent, structured data: purchase orders, invoice approvals, and payment runs, because that is exactly the kind of data automation and AI are best at working with.

Getting that foundation right does more for AI adoption than any tool decision on top of it.

Give your finance admin a foundation AI can actually build on.

Purchase-to-Pay brings structure and visibility to purchasing and approvals, so automation has clean, consistent data to work with.

See Purchase-to-Pay in Action Book a Demo

References
ChatFin (2026) State of AI in Finance 2026: Adoption, Workforce, and the Numbers That Matter. Available at: https://chatfin.ai/blog/state-of-ai-in-finance-2026-adoption-workforce-and-the-numbers-that-matter/ (Accessed: 26 August 2026).
CFO Connect (2026) State of AI in Finance 2026: Adoption Trends, Tools, and CFO Roadmap. Available at: https://www.cfoconnect.eu/resources/reports/state-of-ai-in-finance-2026/ (Accessed: 26 August 2026).