The news that Singapore’s central bank is partnering with Google Cloud highlights the need for the financial services industry to have the right tools – and skills – in using generative AI. Meanwhile, software providers are already crafting solutions so that tools like ChatGPT can be used by treasury professionals.

The Monetary Authority of Singapore’s (MAS) announcement that it would be partnering with Google Cloud on generative artificial intelligence (AI) signals a step change in how tools like ChatGPT are being perceived. Discussions have moved on from the potential of generative AI – which produces natural language text, as well as other media – and now the tools are already being integrated and used by finance professionals.

At the end of May, Singapore’s central bank announced that it would collaborate with Google Cloud to identify use cases and develop generative AI solutions. Also, MAS said in a statement that the two organisations would focus on developing talent and AI skills. Vincent Loy, Assistant Managing Director (Technology) at MAS said, “We hope to inspire greater adoption of responsible generative AI in the financial sector.”

As interest in generative AI increases exponentially, such tools and skills are needed. And as the recent report The 2023 Finance Professional’s Guide to Everything AI, ML and ChatGPT states: “AI will not replace you. Someone using AI will.”

In the course of interviews with treasurers, Treasury Today often hears how AI is being applied, and particular interest seems to lie in cash forecasting, and also in automating reconciliation.

Many companies are exploring how generative AI can be applied to their work environments. There are, however, concerns about how the learning model of open tools like ChatGPT use the information that is provided by users. This is an issue that Samsung already encountered when some of its employees, who were using ChatGPT put sensitive company information – proprietary source code – into the open domain. Since then, the Korean multinational has banned its employees from using generative AI on work devices and is reportedly working on its own in-house AI chatbot tool.

Sharing company information with a third party is often a concern with the finance industry, but there are ways for treasury teams to use these tools without compromising their own data security. Software providers are already developing tools so that their solution functions as a layer between the public AI and the corporate’s financial information.

Technology provider Kyriba, for example, is keeping a close eye on how generative AI can be used by treasury teams. Bob Stark, Global Head of Market Strategy for the company, writes in a recent blog about the potential of such tools. Treasurers, for example, could use the chat tool as part of their regular treasury management system (TMS) and ask it questions like ‘What’s my yen exposure?’ and get an immediate answer, given in natural language. Another use, notes Stark, is in treasury documentation where the generative AI could do the bulk of the writing of manuals and policies and could also keep tabs on what needs to be updated as the risk environment changes. There is also the potential for treasury education whereby the chat tool can give informative answers to questions like ‘When should I hedge?’ or ‘How do I reduce bank fees?’, Stark writes.

One question many have is that ChatGPT doesn’t do mathematics (and at times its calculations can be nonsensical).Many treasurers may be forgiven for thinking that this text-based tool cannot be used for a profession that relies heavily on numbers. This is something that Joseph Drambarean, Chief Product Officer at technology company Trovata, notes in his blog. He explains that the tool his company has developed merges Trovata’s platform – and accompanying number crunching of a corporate’s bank data – with ChatGPT. This, in effect, means that the generative AI has been trained to use the platform on your behalf. Now you can ask it questions like ‘What’s my cash burn rate over the past 30 days?’ and it will give an answer within seconds.

With these tools now being made available, treasurers would do well to upskill and learn how to use them, and in turn – like the collaboration at MAS – they could inspire adoption of responsible generative AI in finance.