How agentic AI is changing finance

AI promised to transform finance. Instead, it’s mostly helped people work faster by summarising reports, analysing data, and answering questions. While that’s extremely useful, it’s not exactly transformational. AI relies on people driving the process, leaving finance teams more productive, but still drowning in manual execution.

Until now.

Agentic AI changes the equation. Rather than waiting for prompts and acting as an assistant, agentic AI can execute routine tasks autonomously, with people overseeing the process when necessary. And that means finance leaders and their teams are finally free to work on strategy, not admin. That might sound great, but there is a caveat. Your agentic AI needs to understand European finance complexity as a foundation, not a feature.

Here’s how agentic AI is changing how finance works, plus the new AI agents Pleo is launching.

Why traditional AI hasn’t transformed finance

Finance teams have embraced AI, but there’s still a gap between enthusiasm and day-to-day adoption. New research from Pleo shows that while 78% of finance decision-makers and finance professionals think their future depends on a deep understanding and use of AI*, 68% say AI skills, training, and confidence are severely lagging*.

Part of the problem is that today’s AI still needs people to drive the process. Someone has to ask the question, review the response, decide on an action, and then complete the work. While AI might speed up the process, it hasn’t fundamentally changed how finance operates.

That means many finance leaders are still tied up with the administrative burden of managing AI, rather than focusing on strategy. But with agentic AI, that’s changing.

From answering prompts to taking action

Agentic AI doesn’t rely on human prompts or step-by-step instructions. Instead, it can make autonomous decisions and then take action.

Imagine an employee pays for a conference ticket. Instead of waiting for them to upload a receipt and submit it for approval, an AI agent can automatically find the receipt in their inbox, categorise the transaction, check it against the company’s expenses policy, and submit it for approval. Finance teams only need to step in if something falls outside the rules.

The same principle applies across all areas of finance. AI agents can reconcile accounts, monitor cash flow, process invoices and flag exceptions. That leaves finance teams free to focus on oversight and strategic direction, not never-ending manual execution.

This represents a shift from workflow to workforce. Rather than simply helping people complete tasks faster, agentic AI becomes an extension of the workforce, handling the routine administrative work that slows finance teams down.

Keeping humans in the loop

Despite the growing potential of agentic AI, this isn’t about replacing finance professionals. As Marija Nakevska, Chief Product and Technology Officer at Pleo, explained: “Pleo’s vision for the future of finance is specialised AI agents that shift finance teams from execution to supervision.”

In practice, AI handles the repetitive work, while finance leaders set policies, review exceptions and make strategic decisions.

Rather than replace finance leaders, agentic AI aims to reduce their administrative burden so they can actually focus on strategy, not micromanagement. “High-level financial strategy will always be something best suited to the experts, and while agentic models can enrich their insights and actions, the responsibility for setting direction will remain with the people using them,” added Marija.

The challenges of agentic AI in Europe

Getting agentic AI right is difficult in Europe. Sixteen markets, multiple currencies, VAT rules that shift by country, and regulations that demand a complete audit trail for every automated action. Finance leaders need to trust what their AI is doing at every step.

So agentic AI in Europe needs to be honest from the start: it can't replace judgment calls. It can only clear the path so leaders can make them faster.

At Pleo, that's exactly what we've built. Our AI agents handle the noise invisibly: receipts get categorised, policies get flagged, anomalies get caught. You see the signal, not the clutter. But when it matters — when judgment is required — we interrupt, immediately, with full context to explain exactly why we stopped.

Our Pleo-provided agent works across Slack, WhatsApp, Microsoft Teams, and SMS, proactively surfacing what you need to decide on. Our MCP interface lets any AI (Claude, ChatGPT, Copilot, etc.) connect directly to Pleo and ask questions. Query spend. Process expenses. Run reports. All from the tools you already use.

Finance leaders don't need AI to replace them. They need AI that knows its limits, works invisibly where it can, and steps back immediately where it matters. That's how you get your time back, not by removing decisions, but by removing everything that's standing between you and making them.

To discover more about AI in the EU, download Pleo’s ‘Europe’s AI spending surge’ report

*This research is taken from a forthcoming report from Pleo, launching in September 2026. The survey was conducted by Sapio Research in April 2026, among finance decision-makers and finance professionals in organisations with 51 to 1,000 employees in Germany, Spain and the UK.