Treasury 2030: How Artificial Intelligence is Transforming the Role of the Treasurer
Introduction
The financial landscape is undergoing a profound transformation, driven by the rapid rise of artificial intelligence (AI). As highlighted during the Agicap Treasury Day in Paris, this technological revolution is often described as the most significant since the advent of electricity—surpassing even the impact of computers, the internet, and smartphones. “We are living in an exciting era with a technological revolution like humanity has never experienced before with artificial intelligence,” opened Mickaël Jordan, CRO of Agicap, setting the stage for a forward-looking discussion on the future of treasury management.
For mid-sized companies, the implications of AI are particularly strategic. These organizations must balance agility and efficiency with the need for robust risk management and compliance. The integration of AI into treasury functions promises not only to automate routine tasks but also to enhance decision-making, improve forecasting accuracy, and foster closer collaboration between finance and technical teams. However, this transformation also raises questions about data governance, the evolving role of the treasurer, and the skills required to thrive in this new environment.
This article distills the key insights from the panel discussion featuring Karine Hegbor de Souza, Director of Treasury and Financing at Longchamp, and Dimitri Letellier, Partner in charge of Data and AI at RSM. It provides CFOs and treasurers of mid-sized companies with a structured analysis of the challenges and opportunities presented by AI, practical recommendations for implementation, and a vision for the treasury function in 2030.
1. Data Responsibility and Collaboration in the Age of AI
1.1. The Shifting Landscape of Data Ownership
The digitalization of financial processes has generated an unprecedented volume of data, creating both opportunities and challenges for treasury teams. As Dimitri Letellier noted, “The trend is not new: as we digitalize and automate flows, technical teams, especially IT, become central to business processes.” This evolution raises a critical question: will finance departments lose control over treasury data to more technical teams?
The answer, according to the panel, lies in proactive engagement. Finance leaders must take the initiative to launch AI-related projects within their own teams, building technical competencies alongside financial expertise. This approach not only preserves the finance function’s ownership of data but also fosters a virtuous cycle of innovation and collaboration with IT.
1.2. Building Synergy Between Finance, Data, and IT
At Longchamp, the treasury team—comprising just three people—manages global cash operations across 26 countries. Karine Hegbor de Souza explained how AI has already enabled her team to work more autonomously, for example by automating the translation of banking contracts in multiple languages. “AI is really about working more efficiently and going further in how we operate,” she emphasized.
Longchamp’s approach is to maintain distinct Data and IT teams, both serving the needs of business units. The Data team’s first mission is to identify and prioritize the requirements of each department, including finance. This ensures that technical innovation is always aligned with business objectives, and that expertise remains distributed rather than concentrated in a single function.
1.3. From Competition to Co-Construction
The panelists agreed that the future lies in reinforced collaboration, not competition, between finance and technical teams. “The idea is not to transfer responsibility to purely technical teams, but to co-construct with IT as a business partner,” said Dimitri Letellier. This collaborative model enables finance to drive innovation while leveraging the technical capabilities of IT and Data teams.
Such an approach is particularly relevant for mid-sized companies, where resources are often limited and cross-functional cooperation is essential for success. By fostering a culture of shared responsibility, organizations can ensure that AI initiatives deliver tangible value without compromising data governance or business alignment.
2. AI-Driven Forecasting: Promise and Practicalities
2.1. The Evolution of Forecasting Tools
One of the most anticipated applications of AI in treasury is the automation and enhancement of cash flow forecasting. The panel discussed the journey from initial enthusiasm—sparked by the public release of tools like ChatGPT—to a more measured assessment of AI’s capabilities. “After the initial hype, many organizations found that real-world use cases were still limited or unreliable,” observed Mickaël Jordan. Indeed, many organizations experienced a “deceptive period,” where the transition from brainstorming and proof-of-concept to real, scalable use cases proved challenging. Early AI models often struggled with the complexity and variability of financial data, and their outputs were not always reliable enough for critical decision-making.
Despite these early hurdles, the landscape is now evolving rapidly. AI is increasingly being used to automate and enhance specific forecasting tasks. For example, models can now accelerate bank reconciliations, a process that once required significant manual effort. In some ERP systems, these reconciliations are now performed almost instantaneously, freeing up valuable time for treasury teams. Furthermore, AI is being deployed to predict payment delays at the invoice level, taking into account a wide range of parameters such as client payment history, the number of unpaid invoices, and average payment cycles. By aggregating these internal data points, AI models can provide more granular and reliable forecasts, enabling treasurers to anticipate cash inflows and outflows with greater precision.
2.2. The Prerequisite: Data Quality and Historical Depth
Despite these advances, the successful deployment of AI in forecasting depends on the availability and quality of historical data. As Karine Hegbor de Souza cautioned, “When we talk about AI and data processing, we must not underestimate the impact and the time it takes to have correct data.” At Longchamp, the integration of subsidiaries into the treasury system has sometimes limited the availability of historical data, necessitating continued reliance on Excel for certain tasks.
For mid-sized companies, this underscores the importance of investing in data management and validation before embarking on AI-driven forecasting projects. Only with a solid data foundation can AI deliver accurate and actionable insights.
2.3. Security and Confidentiality: Managing the Risks
The panel also highlighted the risks associated with open-source AI tools and the need to protect sensitive company data. “It is very important to protect company data and ensure it is not made available to just anyone,” stressed Karine Hegbor de Souza.
The widespread adoption of open-source and public AI tools, such as ChatGPT, can inadvertently expose sensitive data if not properly governed. AI tools can spread quickly within an organization, often without formal oversight, increasing the risk of data leakage—especially when these tools are hosted in regions with less stringent data protection standards.
To mitigate these risks, the panel recommended several concrete actions. First, organizations should implement clear AI usage policies and regularly audit access to external AI platforms. Second, staff must be trained to recognize the importance of data confidentiality and the potential consequences of sharing sensitive information with external tools. Finally, companies should prioritize the use of secure, enterprise-grade AI solutions integrated within their existing SaaS or TMS platforms, which offer better control and compliance.
This is not only a matter of regulatory compliance but also of maintaining competitive advantage and stakeholder trust. As AI becomes more deeply embedded in financial processes, robust data governance will be a non-negotiable requirement.
3. The User Experience of Tomorrow: Augmented SaaS, Not Replacement
3.1. The Future of Treasury Management Systems
A key question for the future is whether AI agents will replace traditional treasury management systems (TMS) and SaaS platforms. While some industry leaders have predicted the “end of SaaS,” the panelists were more circumspect. “We are more likely to see augmented SaaS platforms, supercharged with AI, rather than a rapid replacement by agents,” said Dimitri Letellier.
Current SaaS providers, including Agicap, are already integrating AI features such as automated reporting and natural language data visualization. These enhancements improve usability and efficiency while maintaining the structured environment necessary for reliable financial operations.
3.2. Human Oversight Remains Essential
Even as automation increases, human oversight remains critical. “It is essential to control the data and information provided by AI before making decisions,” insisted Karine Hegbor de Souza. Changes in business activities, evolving strategies, and the inherent limitations of probabilistic models mean that expert review and validation are indispensable.
The risk of over-reliance on AI is real: “If we only rely on AI or the information it provides, we may end up making decisions that are not at all correct,” she warned. The role of the treasurer will continue to require judgment, experience, and a deep understanding of the business context.
3.3. Refocusing on High-Value Activities
By automating routine tasks such as bank reconciliations, AI frees up treasurers to focus on higher-value activities: financing, negotiations with banks, risk analysis, and strategic planning. “If AI can handle the daily problems, that would be very good, because it allows us to devote more time to what really matters,” said Karine Hegbor de Souza.
This shift is particularly beneficial for mid-sized companies, where treasury teams are often small and forced to do multi-tasking. Mastery of digital tools becomes a key differentiator, enabling treasurers to deliver greater value to their organizations.
If you're interested in optimising the structure of your finance team for more effective cash management, whether or not you are already using AI, feel free to download this e-book:
4. Skills and Training: Preparing Teams for the AI Revolution
4.1. Building Technical and Human Capabilities
The consensus among the panelists is clear: ongoing training is essential to ensure that finance teams are not left behind by the AI revolution. “The key is to provide training on technical aspects and risk management, but also to develop deep business knowledge and process improvement skills,” explained Dimitri Letellier.
Experience shows that it is often easier to train finance professionals in technical skills than to teach technical experts the nuances of finance. As AI accelerates the ability of business users to create workflows and automations, a strong foundation in financial processes and strategy will remain the primary source of differentiation.
4.2. Cross-Functional Teams and Continuous Learning
At Longchamp, cross-functional teams have been established to monitor AI innovations and ensure that each department is represented in the development of new tools. “When I started my career, mastering Excel was essential. Today, AI is becoming just as obvious,” said Karine Hegbor de Souza. This mindset is especially important for mid-sized companies, where treasurers often wear multiple hats.
By fostering curiosity, adaptability, and a willingness to engage with new technologies, organizations can ensure that their teams are ready to capitalize on the opportunities presented by AI.
Conclusion
The Agicap Treasury Day panel provided a comprehensive overview of the challenges and opportunities facing treasurers as AI reshapes the profession. Key takeaways include:
Collaboration is key: Finance, Data, and IT must work together to drive innovation while maintaining data governance.
Data quality is foundational: Reliable AI-driven forecasting depends on validated, high-quality historical data.
Augmented SaaS is the future: Rather than replacing existing platforms, AI will enhance them, improving efficiency and user experience.
Human expertise remains vital: Automation must be balanced with expert oversight to ensure sound decision-making.
Continuous training is essential: Teams must develop both technical and business skills to stay ahead of the curve.
For CFOs and treasurers of mid-sized companies, the path forward involves incremental, practical steps: invest in data quality, pilot AI use cases with clear business value, and prioritize training and cross-functional collaboration.
The integration of artificial intelligence into treasury management is not a distant prospect—it is already underway. As one panelist aptly summarized, “AI is the anti-stress for making decisions with peace of mind.” By embracing this transformation, finance leaders can position their organizations for greater agility, resilience, and strategic impact.
Agicap remains committed to supporting mid-sized companies on this journey, providing solutions that combine the best of human expertise and technological innovation.
Watch the replay of the panel discussion (in French).
