AI has quickly proliferated throughout the business world, with the introduction of ChatGPT creating seismic waves across a broad number of industries. This widespread adoption has put the spotlight on AI, with RBC noting that it has turned “from a long-term bet into a near-term reality that has the potential to upend a broad array of industries.”[1] While predictive AI for finance and accounting functions, such as Cash Management, is already somewhat mainstream, we’re still in the nascent stages of how AI can improve these operating models. In this article we’ll dive into the impact of AI on Cash Management, how it’s leading a banking paradigm shift, and the importance of balancing adoption with security.
The impact of AI and machine learning on Cash Management
AI is poised to transform Cash Management, including but not limited to the following functions:
-
Cash flow forecasting:
One of AI’s biggest use cases in Cash Management is around cash flow forecasting, a function that treasurers frequently cite as one of their biggest challenges. Treasury teams are reliant on accurate cash scenario modeling to facilitate strategic decision making and encourage profitability and efficiency. For example, machine learning can be applied to a rich historical dataset to determine trends and patterns to forecast the future cashflows within a defined confidence level or time horizon. When not executed correctly, inaccurate forecasting can lead to problems such as liquidity risk and inadequate working capital management.
-
Reconciliation:
At volume, reconciliation can quickly become both complicated and time intensive. AI for Cash Management can cut down on both errors and time spent by automating steps such as searching for duplications or enabling anomaly detection by using pattern recognition. This functionality can help treasurers focus on tasks where they can really make a difference instead of becoming bogged down in daily payments that can and should be automated.
-
Fraud and cyber prevention:
AI can be a powerful tool to protect teams against cyber-attacks by harnessing the power of data to prevent fraud. This technology can discern minute details such as whether you prefer to scroll or click through a page, and then aggregates this data to create a biometric profile of online behavior. While fraudsters may be able to gain access to your ID and passwords to pretend to be you, it’s significantly more challenging to mimic your keyboarding habits. This provides an extra layer of protection against social engineering scams as any anomalous activities can be identified prior to a transaction being initiated.
How banks can balance AI adoption with security
Adopting AI technology in banking means balancing innovation with security. It involves embracing technology to help treasury teams yet keeping regulatory requirements, privacy, and fairness as top priorities. There are specifics around security, e.g. how AI relies on collecting customer data for its training models, how this data could be compromised by fraudsters or competitors, and implications of potential security breaches on banks from both a reputational and monetary perspective. Bad actors are a major concern in this area, as they have the potential to connect with cloud providers and exploit or use that data model for their own training. Threat actors have also become more sophisticated in their AI-generated phishing email scams. Spotting a phishing email was previously simple due to the multitude of spelling errors or other related indicators.
However, the adoption of AI has produced phishing scam calls, emails, and text messages that are vastly more complex. Threat actors have also started creating AI-generated individual codes that, when combined, can cause a malware attack. To help combat these AI risks, the National Institute of Standards and Technology (NIST) has created its first ever NIST AI Risk Management Framework (RMF), which allows organizations to develop AI systems in a safe, secure, resilient, accountable, and transparent system with enhanced privacy.
How to know if your bank is providing you with the right AI tools
Your bank’s AI tools should be able to transform raw data into actionable next steps, providing you with forecasting scenarios based off historical information and accurate future scenarios. Banks mired by legacy infrastructure may be unable to provide the above if they’re unable to access and interpret their data for the benefit of treasury teams. Cash Management can benefit tremendously from AI. From saving valuable time to increasing security, treasury teams will reap the benefits by adopting this technology into their operating models.
