"Trendy AI" versus "hyper-automation"

Fashionable or trendy moves should come after automation optimization.

Today, AI refers to the adoption of Artificial Intelligence (AI) techniques that are currently popular or in vogue within the industry. As explained, in the context of treasury management, trendy AI might involve using AI algorithms for tasks such as cash flow forecasting, risk management, or fraud detection. However, these AI applications are not (yet) deeply integrated into treasury processes or may be adopted primarily because they are fashionable, rather than because they provide significant value. Furthermore, before adopting AI solutions, some well-advised treasurers could prefer the objective of hyper-automation, because AI techniques remain new, data lakes are not well organized or ready to be used and uniformed or standardized (i.e. on formats).

Hyper-automation is a broader concept that encompasses the use of multiple advanced technologies (including AI), to automate further and streamline business processes comprehensively. I see hyper-automation as an intermediary step toward full automation and AI broad use. There are solutions to optimize your treasury architecture and hyper-automate processes, e.g. KANTOX for FX management automation or FENNECH for reconciliations and many other issues, etc...

In treasury management, hyper-automation involves not only the possible use of AI (it is coming but not yet generalized) but also the integration of robotic process automation (RPA), machine learning, Extract Transform Load (ETL) solutions, and other technologies to automate repetitive tasks and optimize workflows.

Hyper-automation aims to achieve end-to-end automation across various treasury functions, from cash management and liquidity forecasting to compliance and reporting. Hyper-automation focuses on achieving significant operational efficiencies and improvements in treasury processes through a holistic approach to automation, until AI will be able to go even beyond.

In summary, while both trendy AI and hyper-automation involve leveraging advanced technologies like AI in treasury management, the former may imply a more superficial or selective adoption of AI techniques (some solutions do not exist yet – are under construction/development), whereas the latter represents a comprehensive and strategic approach to process automation and optimization.

AI everywhere… but not now…

Even if we all know that AI will change the world, it may not be immediately in treasury. We must be aware that the euphoria surrounding AI will continue and that great solutions will emerge for treasury. But treasurers should be part of the co-creation of solutions to indicate banks what they need and how AI could help. AI everywhere is coming but not yet there, I guess. For example, I do believe in terms of decision-making solutions, it will certainly help in the FX management to optimize strategies within predefined boundaries in polices. AI could recommend increase or decrease in hedging within a corridor of hedging objective fixed by policy. In working capital need management also, we can expect AI to become a mean to determine how to optimize working capital needs (WCN) as well as in analytics around working capital factors or how to optimize WACC, is another example. Regarding WCN, AI will analyze figures and detect changes and sources of changes by analyzing ratios and KPI’s. Detection of trends will help correcting shifts if any. Eventually, for predictive cash-flow forecasting, we can use figures of receivables to detect changes in behaviors from customers and to detect deterioration of customer solvency and creditworthiness (e.g. AIvidens).

The best of AI is yet to come.

Treasurers and their IT providers should continue exploring where AI can help management optimization and enhancing internal controls for better efficiency. The best of AI is yet to come. Treasurers can be part of the solution and play a key role in defining what to develop, in cocreation with suppliers and banks. In the meantime, the next level of automation can be pursued in a variety of ways, as a kind of intermediate step before AI goes mainstream. But let's be reassured, at the end of the day, AI will not replace the treasurer but will allow him or her to focus on analysis and action, a more strategic role than ever and decision support, where hyper-automation helps to spend less time doing repetitive, manual, or semi-annual tasks and reduce what the machine can do better and more efficiently, wherever that may be.

AI smarter than human?

We don't know what the future of AI (which goes far beyond a mere fashion trend) will bring. If we take Elon Musk's word for it: "My guess is that we'll have AI that is smarter than any one human probably around the end of next year", it's chilling. According to OpenAI & Meta, they are working on AI models that are prepared and skilled in reasoning, which would give the machine the ability to fix "hard problems". The human 'treasurer' has his/her place. On the other hand, the fact that treasury teams are often under-staffed will give them some much-needed breathing space and capacity for reflection and analysis. The agile treasurer will adapt and be in the cockpit of a more efficient machine, which will finally become a decision-making tool and not just a simple reporting tool. The role will evolve upwards, with more thinking and less execution. But whatever the timing, the path has been mapped out, and it will first involve hyper-automation, and later move towards artificial intelligence, the ultimate level of decision support. We're still a long way from having seen everything when it comes to corporate treasury, and I'm delighted about that.

Fran?ois Masquelier, CEO of Simply Treasury – Luxembourg April 2024.