While most of the panelists saw AI as an “enabler” for the improving the use of supply chain finance, there were questions raised about the extent of its potential. Audience members were equally vocal about their concerns around the accuracy of AI and how long it might take to develop fully-AI enabled solutions.
Christophe Juvanon, head of solutions consulting at Taulia, a fintech which launched an AI-supported cash forecasting tool last month, said: “AI is an enabler to do financing further in the supply chain.
“A cash manager or treasurer should have the tools to predict their cash position over three months. AI will give actionable intelligence to manage your cash in your supply chain.”
He explained that AI can allow companies to use vast amounts of data gathered on supplier or buyer behaviour to optimize their SCF programmes.
“This is where we are today and moving to scaling this – looking at buyer payables and receivables and working out some formulas,” he said. He added that as momentum grows, AI will be able to harness more data to analyse the risk on unapproved invoices and purchase orders, enabling more pre-shipment and purchase order financing. This is starting to be explored but remains limited.
Philipp Schoenbucher, co-founder and chief data scientist at Previse, said that it was important to consider what you want to achieve from using AI rather than see it as the answer to all business requirements.
“Look at what you are fundamentally trying to achieve and where you can add value in the real world. If AI as a tool is useful, use it or if not – use something,” he said. Previse is a start-up that uses algorithms to analyse the data from invoices and predict which ones might be problematic and pays the rest instantly.
Laurent Tabouelle, COO of software provider Codix Group, was the voice of dissent on the panel, noting that AI should not be viewed as “one magical solution for everything.”
He explained that AI as a machine learning tool could be very reliable in helping companies decide – for example – if a buyer is a “good” or “bad” payer, but that notion of good or bad is created by a human having input information into the system which will take time to build up.
He refers to AI SCF-related solutions currently being explored in the market as more of an “augmented decision-making” tool, rather than “pure AI”.
He said developing certain AI-enhanced solutions would be a “slow process” and that “we are ages away from what we can do.”
In contrast, Bart Ras, visiting fellow, Windesheim University of Applied Sciences, told delegates that AI would play an important role in shaping the future of SCF, it was just up to you as a company whether you want to use it.
“Let’s stop kidding ourselves – it is [AI] around. Do you use it for your benefit? Or ignore it,” he asked.
He told delegates that if they were keen to develop AI-supported products they should look to industries beyond SCF.
“Look at other industries that have a head-start in this field and see what you can adopt. And hire some people have nothing to do with your current industry but worked with this [AI] elsewhere,” he said.
“If you look at what quantum computing can do now, you will realise the future is a lot nearer than lot of people think,” he added.