Rather than introduce sweeping AI strategies, corporate treasurers are using AI to solve specific pain points in a modular approach.
Take a utility company that is battling to complete last mile automation in reconciliations, suggests Tibor Bartels, Managing Director, Head of Transaction Services Americas at ING who works across cash and liquidity management, working capital and ING’s trade offering. AI can help optimise reconciliations to support the company further reduce human interaction in its reconciliation processes.
In contrast, a mining company which has a much lower level of transactions but is working in an industry buffeted by commodity prices, geopolitics and sanction risk, would use AI for forecasting. “AI is not a light switch you turn on and say everything is AI-driven. It’s a gradual process that should be tailored to the needs of a specific organisation,” Bartels tells Treasury Today.
Moreover, modular cash forecasting and reconciliation strategies are being fanned by treasury’s quest for certainty amidst growing geopolitical instability. And forecasting is one of the most exciting areas of innovation because it can be applied to everything from predictions on commodity prices and exchange rates, to future sales. Successful strategies depend on access to data to feed traditional AI forecasting models and drive accuracy going forward, says Bartels.
In another trend, companies are also leaning into GenAI (distinct from traditional AI because it learns from data to generate original content) in the same modular approach to support KYC and onboarding. The enduringly challenging process – that also comes with regulatory restrictions and complexities around legacy ERP and other treasury tech –can take months and becomes even more laborious when corporates seek multiple bank relationships.
ING has now scaled a GenAI model to support onboarding that is currently being used by 200 clients. The data KYC requires is typically already held by companies and scattered through corporate mailboxes or credit systems. The bank uses GenAI to bring that data into one place and highlight potential KYC risks. “We are using GenAI to answer KYC questions and assess the risk levels. The early results show dramatic time savings and productivity gains,” says Bahadir Yilmaz, ING’s Chief Analytics Officer.
ING is also creating structured data from different jurisdictions to provide better insights for clients. The bank has a footprint across 100 countries with a myriad of different legal frameworks, data recording processes and data formats. ING is now drawing that data into one single model to support clients with a holistic view “of what is going on,” says Yilmaz.
“AI is only as successful as the data sources we have,” Bartels adds. “The more historic and regulatory data we can gather enables our clients to be more successful in the regions we do business.”
Yilmaz and Bartels observe treasury teams have budget to spend when it comes to AI. Companies are prioritising spending on the technological tools that can support increased efficiency and fraud prevention, ease onboarding or the integration of new technology and products like virtual cash management over other strategies. “When it comes to fast-changing technology, price is not the driving factor. Corporates want a strategic driver to optimise processes and keep them secure in a rocky future,” says Bartels.
Perhaps the most important first step to integrating AI is to begin by clearly distinguishing between GenAI and AI. “People think GenAI should be applied to every business problem, but it’s not the case. For example, if treasury is seeking insights on the credit worthiness of a client, GenAI won’t add much value,” says Yilmaz.
An AI strategy requires deep strategic thinking and cross functional teams spanning product managers, researchers and data scientists. Yilmaz and Bartels also urge corporates to upskill teams and hire from the market. “There is a human transformation element to the story. If corporates want to do more with AI, they need a transformational element.”
