Treasury tech is getting a long-awaited upgrade

How can treasurers build a more effective tech stack?

“Part of the solve is to have the right decision-makers all at the table to ascertain what is the best tech solution going forward, what is really needed and what can be eliminated,” says Jeff Norfolk, Product Marketing Strategy Lead at J.P. Morgan.

Those decision-makers should include people from other internal teams such as sales and operations, but also third-party software vendors and consultants who can recommend best practices.

They may recommend a better TMS that can integrate more technologies into a single platform. TMS tools are a key trend in this space. However, even the best TMS has some integration limitations. For instance, none of these tools can undertake effective advanced cash forecasting methods.

What else can help?

One challenge in integrating tech stacks is that each component can generate data in different formats and locations. Tools that can gather disparate data into a centralized repository are therefore valuable.

“That’s where application programming interfaces (APIs), robotic process automation (RPA) and middleware comes in,” says Alberto Hernandez Martinez, Industry Solutions at J.P. Morgan. “APIs enable disparate systems to interact with one another, whereas RPA involves software bots that automate the transfer of data across various tools, and middleware acts as a connector between diverse technologies.

The more that APIs, RPAs and middleware can integrate tech stacks, the more AI can do for treasurers. AI-based virtual assistants could have all the necessary data at their disposal to perform complex analysis and take logical next steps, while providing a simple natural language interface for the user.

“AI technologies, such as the Model Context Protocol (MCP), that enable large language models (LLMs) to securely connect to tools, data sources and other LLMs, or AI agent-to-AI agent technologies that allow AI agents to securely communicate with each other will be a step towards a more AI-enabled treasury function,” says Hernandez Martinez.

As is often the case, compliance and risk management will be key considerations. Companies would be remiss to have an AI agent complete a $50 million transaction without human sign-off, so these solutions will need to balance automation with human oversight. Hernandez Martinez adds, “While we don’t know how many years it will be until AI treasury assistants are mainstream, we do know AI agents are being developed and deployed now. If treasurers prioritize the tools and collaboration needed to create an efficient tech stack, they’ll be in prime position to adopt them when they arrive.”