We have all, no doubt, heard of ChatGPT, which uses natural language processing to engage in conversations, answer questions and generate content and is used by many businesses for tasks such as answering complex business questions.

Now meet Treasury GPT, which according to fintech FIS, uses machine learning (a subset of artificial intelligence that automatically enables a machine or system to learn and improve from experience) to provide users with high quality, informed guidance, enabling them to focus less on low value administrative tasks and more on liquidity management and strategy.

The tool sits inside the company’s TMS and uses Microsoft’s Azure OpenAI Service. According to Jacqueline O’Flanagan, Head of Financial Services for the Americas at Microsoft, it addresses a host of efficiency and workflow challenges for corporate treasurers and CFOs.

“Treasury GPT is certain to bring immediate productivity and empowerment benefits to its users and I look forward to seeing how it evolves,” says Kevin Permenter, Senior Research Director Financial Applications at market intelligence firm IDC.

Potential core applications of AI in treasury include:

• Predictive cash flow forecasting where AI analyses historical transaction patterns and external factors – such as interest rates or currency fluctuations – to predict future cash flows.

• Liquidity management and optimisation, where AI models optimise working capital by suggesting the best cash allocation across accounts.

• Fraud detection and anomaly identification, where AI continuously monitors transactions and flags suspicious activity in real time so if an unusual payment is detected, the system can pause the transaction and request human verification.

• AI-powered virtual treasury assistants, where chatbots and AI assistants provide on-demand insights, answering questions such as ‘what is my projected cash balance next quarter?’ or ‘are there any anomalies in our payment flows?’ and making treasury management more intuitive and accessible.

“While APIs and AI offer immense potential, the successful implementation of these technologies hinges on several key factors that treasury units need to consider,” observes Ulrika Claesson, Head of Open Banking Solutions at Nordea. “Firstly, the maturity of ERP and treasury systems – how well can they handle and process real time data? Secondly, data quality and security – how do businesses ensure data integrity while mitigating cybersecurity risks? Thirdly, human oversight – AI is a powerful tool, but human judgment remains critical for decision-making.”

Alberto Hernandez-Martinez, Executive Director Industry Solutions at J.P. Morgan agrees that AI has introduced a new era of precision, efficiency and strategic insight to cash and liquidity management.

He suggests AI-powered forecasting models can reduce error rates by up to 50% compared to traditional methods according to case studies from multinational corporations.

“However, users of these complex models should focus on ensuring the results they generate can be clearly explained and easily interpreted, particularly in regulated financial environments,” adds Hernandez-Martinez.

He also observes that successful implementation of AI in treasury management requires more than just technological adoption. It demands a shift in mindset, a commitment to data quality and a willingness to blend human expertise with machine intelligence.

“Instead of replacing skilled treasurers, these powerful tools enhance treasury’s strategic impact, turning financial insights into high level executive decisions,” he notes. “With AI and human intelligence working in tandem, the partnership transforms liquidity management from a support function into a competitive advantage.”