The future of treasury: How AI is transforming treasury management

Applications of artificial intelligence are accelerating in the financial world. Not only is AI starting to have a meaningful impact on corporate treasury operations today, but forecasters expect more exciting, transformative applications in the future.

Given the growing emergence of AI in corporate finance, how should treasurers respond?

“The first thing is not to fear it,” says Vipul Kaushal, senior vice president and group head of the Digital Transformation Team within U.S. Bank Global Treasury Management.

AI offers corporations the ability to do more with less as they grow. This naturally creates concern about the impact AI might have on current treasury jobs. However, Kaushal says he expects the use of AI in treasury management to lead to new and higher-level positions. “There still needs to be a human to train the bot on the logic that needs to be applied in a treasury application,” he points out.

Overcoming any fears about AI in treasury management and embracing the technology is wise, Kaushal says, and the first step for treasurers — one that many have already taken — is to look for ways AI can make their people more efficient. So, in reviewing current and future applications of treasury AI, let’s start there.

More efficiency through intelligent automation

Historically, payables and receivables processing and associated reporting require employees to spend considerable time manually performing predictable and/or routine low-value tasks — such as converting paper records to digital formats, flagging discrepancies and generating routine reports — as well as managing exceptions and disputes. AI can enhance the effectiveness of these tasks while greatly increasing efficiency by reducing time demands. This is where AI has made the biggest impact to date, and where companies should start when leveraging treasury AI.

To explore the potential of AI in treasury operations, Kaushal suggests treasurers begin by identifying core processes in their units, pinpointing inefficiencies in those processes, and then determining the related direct and indirect costs. “From there, they can look for ways that AI technology can enhance or assume those tasks,” he says.

By using basic AI tools, many treasury employees can have their repetitive and manually intensive tasks shifted to machines and their time reallocated to higher-value analytical work.

One particularly promising application for AI in treasury management and payments is account reconciliation. The technology can use data extracted from multiple departmental and accounting sources, matching it against bank statements and the general ledger to automatically reconcile transactions.

Additionally, Kaushal provides examples of intelligent automation solutions that banks are beginning to introduce in service and operations to increase efficiency and improve the treasury customer experience:

• Guidance on selecting payment types. An AI tool can analyze information such as the size of a particular payment, how fast it must be delivered, and the payment types the supplier accepts, and then recommend the most appropriate payment type to use. For example, if Accounts Payable needs to deliver a payment today and the supplier accepts payments via the RTP^® network or some other instant payment method, that instant payment type would be the likely recommendation. If the treasurer has three days to make the payment, the system might suggest ACH or some other method.
• Intelligent email automation to support faster customer service. Intelligent bots that banks are beginning to deploy are reducing response time on customer service inquiries. The bots review incoming customer service emails, assign the inquiries to the appropriate bankers, and even suggest responses.
• Chat-enabled treasury product education. U.S. Bank has scores of treasury products. Kaushal says the bank is developing an AI-supported chat capability that will allow treasury sales team members to ask for information about specific products, get quick responses and reduce research time as they evaluate those products for possible recommendation to a client. The plan is to eventually make chat-enabled product education directly available to bank clients.
  

Cash forecasting support

With the United States lagging in establishing a legal framework around artificial intelligence, banks’ liability concerns are slowing down the release of certain types of AI-enabled solutions. One area where more banks are expected to offer treasury AI solutions, once guardrails are in place, is cash forecasting.

Sub-fields of AI such as machine learning, data mining and advanced statistical modeling can be combined to produce predictive analytics. Kaushal says AI-supported cash forecasting tools will enable treasurers to predict cash flows, so they can plan for investments or loans and better manage their cash.

“Today treasury managers rely on reports we generate for them and Excel spreadsheets they maintain on their own,” he says. “But imagine if they can go into a digital portal to access more real-time data to learn not only how much cash is going out today, but also a month later.”

For example, a business required to make a lot of tax payments could use an AI tool to predict future tax payment outflows. “This is something AI can tell them months in advance so they can plan to have the necessary deposits on hand,” Kaushal says.
 

Managing risk, preventing fraud

AI in treasury can also play an important risk management role.

For instance, U.S. Bank is using machine learning along with optical character recognition (OCR) to digitize business client signatory information and improve the sanctions screening process.

Machine-learning custom logic enables the bank to pull designated data from a client’s signer information documents, digitize that data and structure it chronologically in a database. “As soon as the signatory documents come in, the data is stored digitally with no manual effort and fed into sanctions screening in a real-time manner,” Kaushal explains. “This allows sanctions screening to be completed faster and more securely.”

In treasury operations, AI is also at the center of efforts to fight payments fraud.

One of the best examples of an organization using AI to reduce fraud is the U.S. government. The U.S. Department of Treasury has credited AI-powered tools with helping to prevent and recover over $4 billion in fraudulent and improper payments in fiscal 2024. Recovery of about one-quarter of those funds came from using machine learning to catch check fraud, the government says.

Using AI in treasury management to help clients combat fraud “is something we as a bank are looking to explore more in the near future,” Kaushal says.

The next stage: generative AI

The use of AI in treasury management has progressed in stages — from rules-based automation to machine learning applications to predictive analytics. The next stage will involve treasury harnessing generative AI, Kaushal says.

Generative AI uses machine learning models to learn from large amounts of data. It can produce “contextual responses” to complex questions. Kaushal suggests this higher level of AI, still in its experimental phase, could initially impact treasurers when their banks start using it to “sell solutions in a more intelligent manner.”

In the future, he envisions a bank’s treasury salesperson being able to feed information about a client’s treasury challenges into generative AI. The system will use that information — and what it already knows about the client and its industry — to recommend a combination of products, along with suggested payment methods, file formats and the best way to connect to the bank. In other words, generative AI will be able to detail a holistic treasury solution that the salesperson can recommend.
 

The future of AI and corporate treasury technology

To begin transforming finance for the AI-enabled future, Kaushal suggests treasurers focus on forward compatibility when making technology acquisitions, and ensuring new platforms allow for a future application of AI.

New technology such as AI can help the treasury function move beyond its traditional role as a cost center, add top-line value and strengthen the bottom line.

“At U.S. Bank, we understand this. We help bridge the gap by building tools that make it possible for our clients to take advantage of new technologies and make it easy for them to grow their businesses,” Kaushal says. “We are investing heavily in AI — we’ve established a Center of Excellence around it — but at the same time we’re focused on being very responsible about its use.”

Contact us to learn more about how AI can enhance your treasury department today and tomorrow.