Spendesk’s CEO Shares Vision for AI-Driven Finance

Across financial services, AI is rapidly shifting finance from a backward-looking function into a real-time, decision-making engine. 

Static reporting cycles and delayed insights are giving way to continuous intelligence, where data is analysed and acted upon at the moment decisions are made. 

For CFOs, this marks a fundamental transformation: from overseeing historical performance to orchestrating dynamic, forward-looking strategy.

At the centre of this shift are platforms embedding AI directly into financial workflows.

Axel Demazy, CEO of Spendesk, is positioning the company at the forefront of this evolution. 

With a background in scaling technology organisations and leading data transformation, he sees AI not as an add-on, but as the foundation of a new operating model for finance. 

Under his leadership, Spendesk is moving beyond spend management to become what he describes as the intelligence layer for the modern Office of the CFO.

In this Q&A with FinTech Magazine, Axel delves into how AI is reshaping finance and what it takes for CFOs to build truly AI-native functions.

What does AI-native finance actually look like in practice, beyond automation and copilots?

AI-native finance isn’t about layering tools on top of existing processes, it’s about rebuilding the function around continuous intelligence. 

Instead of static reporting cycles, autonomous agents monitor spend in real time, surface risks early, enforce policy rules and proactively recommend or trigger actions within clear guardrails. 

The result is a shift from retrospective analysis to always-on financial oversight.

A big step in that direction is the launch of Spendesk AI Connect at Money 20/20 this year.

It goes beyond automation, allowing finance teams to interact with their spend data through AI assistants like Claude or Dust, asking questions in plain language and getting instant answers without having to build reports or scroll through dashboards. It turns something that used to be rigid and report-heavy into a conversational experience embedded in the tools people already use.

More broadly, it signals a move from static reporting to real-time, AI-driven access to financial insight, where finance teams can get answers in seconds while still keeping full control over approvals and payments.

How has embedding more than 150 internal AI agents changed decision-making for teams at Spendesk?

The honest answer is that it changed the way we work before it changed the way we decide. 

In 2025, we deployed an AI solution that allows teams to build and run their own agents, to every single employee from day one, without restrictions and without a top-down mandate.

The idea was simple: let every team identify their own pain points and build solutions for them. In total, 213 agents were created, and 135 are now actively running in production.

The concrete impact has been significant across the business. To give one example, our Customer Success team now prepares quarterly business reviews in three minutes instead of two hours, something that has genuinely changed how systematically they engage with clients.

What the numbers don’t fully capture, though, is the cultural shift that came with it.

The best ideas didn’t need to come from the top, often they came from the people closest to the work. 

Today, around 25% of our employees have built their own agent and teams now instinctively ask ‘could an agent handle this?’ before defaulting to a manual process.

That change in reflex is, in many ways, more important than any individual use case

What cultural or organisational shifts are required for finance teams to adopt AI at scale?

Often the biggest barrier to AI adoption isn’t technology, it’s mindset. 

Finance teams now need to shift from control through manual oversight to control through intelligent systems. But of course, that will require trust, built through transparency, experimentation and clear guardrails.

Organisations also need to embrace the ‘ready-willing-able’ model. Teams must be ready to challenge assumptions, willing to evolve their roles and able to learn through hands-on experimentation. 

We’ll see roles evolve too. A new profile is emerging, the Finance Engineer, who layers automation, AI tools and workflow design onto traditional finance expertise.  Analysts become AI orchestrators, controllers become policy designers and Financial Planning & Analysis (FP&A) teams become scenario strategists. 

Where do you see the biggest gap between how companies are experimenting with AI and where they need to be?

Recent Forrester research shows that only a small minority of agentic AI projects move beyond proof-of-concepts, which says a lot about the gap between experimentation and real operational change. 

Organisations are trying out copilots, chat interfaces or small automations without rethinking the underlying system they sit on. 

AI-native organisations don’t bolt intelligence onto old processes. They rebuild those processes end-to-end so AI can operate as part of the structure, not an add-on.

It’s not until companies make that shift, they’ll struggle to unlock the ROI they’re aiming for.

Another gap is governance. Many companies lack clear frameworks for AI accountability, risk management or measurement. 

Without this, AI remains a productivity booster rather than a strategic engine.

How should CFOs rethink control, risk and accountability in an agent-driven operating model?

In an agent-driven model, CFOs must shift their focus from controlling every action to oversight of the mechanisms that govern autonomous agents. Control then becomes more about designing the system than executing the process.

Risk management becomes continuous. Agents monitor transactions in real time, flag anomalies instantly and enforce policies automatically.

This reduces human error but requires CFOs to invest in transparent audit trails and clear escalation paths.

Accountability remains human. CFOs are responsible for judgment, policy and governance – agents are responsible for execution within defined boundaries. 

The CFO’s role evolves into architect, steward and validator of an intelligent financial ecosystem.

What role will SP&A play in reshaping the Office of the CFO compared to traditional FP&A?

For us, Strategic Planning & Analysis (SP&A) represents the next evolution of FP&A in an AI-native world. 

Traditional FP&A focuses on reporting, budgeting and variance analysis. SP&A, on the other hand, focuses on real-time strategy, scenario modelling and decision enablement.

With AI handling data consolidation and forecasting, SP&A teams can spend their time evaluating strategic options, running simulations and advising the business on trade-offs. 

They help bring operations, finance and leadership onto the same page, turning data into decisions the whole organisation can act on.

In this model, SP&A becomes the strategic brain of the CFO’s office, guiding decisions, shaping priorities and ensuring the organisation moves with speed and clarity.

Based on Spendesk’s own transformation, what are the first concrete steps for finance leaders aiming to become AI-native?

The first step is mapping your workflows end to end and identifying where decisions are actually made. AI is most powerful when it’s embedded at those decision points, not bolted onto existing processes.

From there, build a handful of high-impact, low-risk agents, things like anomaly detection or policy enforcement.

These quick wins matter because they build trust. And don’t wait for perfect training programmes, AI literacy comes from doing so create your first agent, experiment, fail, try again. 

You also need strong data foundations and a cross-functional group to define governance and guardrails. 

But, above all, leaders must champion the mindset shift.