While rapid developments in technology, AI and system connectivity are driving fundamental improvements for potential information flows within organisations, a collaborative environment is an essential prerequisite for ensuring this data is optimised to deliver fully integrated dynamic, real-time cash forecasting solutions.
Speaking at the ACT’s annual Cash Management Conference, Dr Jacques Yana Mbena (picutred above), senior vice president solutions architect at cloud-based software and service provider TIS, discussed how technology, including AI, could be used to optimise information flow for real-time cash forecasting. “It is extremely important in the context of digital transformation in the time of AI that skill sets need to be there – not only for using forecasting, but questioning the compliance logic behind this and how you trust what comes out of the system,” he said. “For every treasurer, the nightmare is you have data that you do not trust, that’s effectively worthless for making a financial decision with.”
Mbena outlined the importance of integrating real-time external market data analysis with internal data to inform forecasting logic, and the vital importance of being able to link those data layers: “You cannot be in a silo in your organisation. You need to be more collaborative to capture insight from where the data is coming and then you have to link those data.”
Andrew Griffiths, assistant treasurer at Anglian Water, agreed with the value of collaboration. He cited the vast amount of data that his organisation has and the subsequent challenges of using that dataset for cash forecasting. “There's a journey that any organisation needs to go on in terms of their underlying systems and data sets to make sure they have a robust dataset. We are, as an organisation, on that journey but we’re not quite there yet,” he said. “We have to bring them together, using people and processes rather than machines. But I do think that having that common data set, being able to analyse the data set in a collaborative way is going to be helpful for cash forecast.”
For every treasurer, the nightmare is you have data that you do not trust, that’s effectively worthless for making a financial decision with
Griffiths acknowledged the opportunities and challenges presented by AI in terms of developing cash forecasting. “It’s definitely the direction that things are going in,” he said. However, he believed that to get to that future state “there needs to be a step change. There needs to be more trust in data.”
It requires a better agreed dataset that everybody buys into, Griffiths said, as well as the need for people to develop and change with it. “There’s a reliance on experience and on looking at a set of numbers and whether they pass the ‘sniff test’,” he said. “But, sometimes those personal views can be wrong and you can be led down the wrong path.”
When people get comfortable with the data, and when it is becoming more reliable, “[AI] will start to replace some of the techniques and judgements we have now”, Griffiths concluded. While AI may not currently be capable of using all available data to provide a robust, fast and dynamic cash forecast that treasurers can trust, AI systems will ultimately be able to learn and provide you with the optimum pipe for cash flow forecasting data. “For the rest, it’s up to you whether you are happy to take decisions based on the AI for cash flow forecast,” Mbena said.
AI was revisited in a later conference session around navigating agility in today's business environment. Richard Ransom and Kevin Grant from Bottomline covered how today’s fast-developing business landscape – with economic and regulatory uncertainty as well as shifting market dynamics – required greater agility when managing working capital.
As well as highlighting how cloud-based integrated payment management solutions can support agility, they acknowledged that AI applications, while creating a buzz, were something to be considered purely on their effectiveness and applicability for a particular business or solution requirement. “Make sure that you understand the use case that you’re working to solve; everything you buy technology-wise and the information that technology presents to you must help you make better decisions. The AI model is never accountable for a mistake,” Grant said. “It either works for you or it doesn’t – you know your own use case.”
Phil Lattimore is a freelance journalist
You may also like
