What’s next in AI spend control?

Finance leaders are already committed to AI. Pleo’s latest European Finance Index shows that UK businesses are now investing in enterprise subscriptions and API usage at scale. OpenAI (ChatGPT) entered the UK’s top 20 merchants for the first time, while Anthropic (Claude) climbed from #12 to #7.

The message is clear. AI has moved out of the experimentation stage and is now an embedded part of modern finance workflows.

But CFOs now face another challenge: the shift toward agentic automation in finance. AI isn’t just supporting finance teams; it’s now starting to complete routine tasks on their behalf.

That means finance leaders need to decide not only which agentic AI tools to adopt, but they also need to consider how to govern them. The latest phase of AI in spend control is all about balancing automation with oversight.

From AI adoption to AI governance 

The first wave of AI helped change how finance teams work with information. Our past research shows that CFOs and their teams used it to create financial summaries, analyse real-time data, help with financial planning, and quickly consolidate large amounts of information.

While these capabilities boost productivity and speed up workflows, they still require significant human oversight. Someone still needs to review expenses, chase missing receipts, answer questions about expense policies, and approve payments.

That’s where agentic AI comes in.

Instead of summarising reports and analysing data, specialised AI agents are changing how finance works, executing routine tasks independently and only escalating exceptions that require human oversight. This shift means finance leaders need to ensure the autonomous AI tools they’re relying on are compliant, auditable, and aligned with company policy.

Truth matters more than autonomy 

A lot of conversations about agentic AI focus on how it works autonomously. But for finance leaders, trust is also a crucial part of the equation. Every decision made by agentic AI needs to be accurate, explainable, and fully compliant.

For businesses operating across Europe, that can be complicated to achieve, since you’re working across multiple markets with country-specific rules and regulations. Every automated decision also needs a clear audit trail.

At the same time, finance teams are already under pressure to do more with less. Given that finance teams also spend an average of 125 hours per year switching between disconnected tools, the future of AI in spend control shouldn’t introduce another platform to manage.

The answer is layering intelligent automation onto the tools and workflows you’re already using.

In practice: Agentic AI in spend control 

Vague promises of agentic AI’s potential don’t help CFOs become more productive or deal with regulatory pressure. What’s needed is genuine proof of tools that actually work.

At Pleo, we’re committed to helping finance teams transform spend control. That means we’ve been developing a suite of AI agents designed to handle the noise invisibly.

We’ve already launched our Policy Agent, helping enforce company spend policies in real time and flagging exceptions before they become problems. Soon, we’re launching a Pleo-provided AI agent that works across your existing tech stack to locate missing receipts and keep routine workflows moving.

Our AI roadmap includes the launch of another four specialised agents across expenses, accounts payable, treasury and accounting. Together, they’ll automate routine tasks like capturing receipts, coding transactions, processing invoices, scheduling payments and supporting month-end close. But we’re building something that knows exactly when to escalate decisions and exceptions that need human oversight.

You get time back. Your auditor understands every decision. And, your team trusts it because it's honest about its limits, not confident about its superpowers.

Ready to see Pleo’s AI in action? Book your demo today.