Generative AI: What does it mean for payments?

Generative AI is probably the most talked-about technology of 2023 and as such it has caught the attention of technology leaders at financial institutions around the world. But what does it mean for payments services?

Due to the critical importance of payments systems, these are subject to considerable handling and regulation, covering financial, operational, and overall business risk and resiliency. It is therefore imperative that the implementation of AI technology is carried out responsibly and in compliance with regulatory direction to protect the interests of both the financial institution and its customers.

AI has actually been used in payment processing for several years. However, until now it has typically been embedded within solutions and has not always received much focus. Generative AI platforms such as ChatGPT have made AI tangible, revolutionized the perception of AI, and enabled users to leverage it without the need for coding skills. As such, generative AI is expected to open the door for new payments use cases.

Gen AI and payments: what are the key possible use cases?

AI use cases in payments have the potential to significantly improve regulatory compliance, anti-money laundering and payment processing. They could also create efficiencies for enhancing customer experience and fraud detection, thereby benefiting financial consumers and market participants.

But while there is plenty of focus on the potential applications of tools like ChatGPT, these large language models are just one portion of generative AI – and where payments are concerned, opportunities such as the ability to generate synthetic data are just as significant.

Some of the most promising payments-related use cases include:

• Improving payment initiation. Sophisticated payment initiation solutions can leverage generative AI to extract payment information from invoices and other documents automatically – thereby reducing manual tasks, eliminating errors, and improving the payment initiation stage, before the payment is approved to be processed.
• Fraud detection and prevention. Generative AI could be used to develop more effective fraud detection and prevention systems. By analyzing payments data about past fraudulent transactions, generative AI could learn to identify patterns and develop predictive models to flag suspicious activity in real-time, which becomes more significant in the new era of instant payments.
• AI-powered chatbots. Generative AI chatbots can provide personalized real-time responses, enhancing customer interactions with payment product documentation.
• Automated payment reports. Gen AI could be used to automate various business and operational reports, improving operational efficiencies, while reducing errors and inaccuracies for improving the bottom-line payment business performance.
• Enriching data into a structured format. AI can also play a role in breaking a four-line free text address into a structured format. While the migration of ISO 20022 will require banks’ payments engines to adopt structured data, such a capability can assist with the integration to legacy bank systems.

Paving the way for practical uses of AI in payments

At the recent Sibos 2023 event that took place in September in Toronto, Finastra showcased and demonstrated how AI and Generative AI are explored and have already been embedded in its payments solutions and services. The following use cases were demonstrated:

• Unlocking payments insights with AI-Powered Dashboards. Finastra Payments Insight Dashboards leverage historical payment data to provide actionable insights. These dashboards help improve business processes, reduce costs, and increase revenue. Using machine learning logistic regression classification algorithms, the dashboards unveil hidden factors such as payment source and method, which in turn impact non-straight-through processing rates and payment processing duration. Insight Dashboards are part of our Payments Reports and Analytics add-on offering and are available to our Payments hub customers.
• AI-powered chatbot for payments product knowledge. From the FedNow business guide to Finastra Payments product documentation, Finastra provides a wealth of information designed to support clients in their search for information. Our Gen AI-based Payments Chatbot combines this corpus of information with large language models to bring real-time Q&A to all Finastra payments products users, helping them to reduce their learning time and optimize their payment processing performance.
• AI-powered invoice-based payment initiation. Paper invoices and email attachments are still widely used by billers to get paid by their customers. This typically involves typing bank transfer details from scratch, which can cause payment delays and errors. Powered by advanced Gen AI, our solution automatically extracts payment information from invoices, eliminating the need for manual data entry. Advanced machine learning algorithms are used to identify and correct errors in payment data. The solution automates many of the manual tasks involved in payment processing, such as matching invoices to payments and reconciling accounts.
• Finastra Compliance-as-a-Service. Our SaaS-based pre-integrated payment processing and regulatory compliance solution supports instant payments regulatory compliance with real-time sanction screening and specialized AI/ML-based transaction monitoring. It offers simple integration with new payment networks like FedNow and TIPS for the US and European financial institutions. It also scrutinizes transactions against multiple sanction lists to minimize false positives and ensure compliance with evolving regulatory laws.

The way forward

The future for leveraging AI in payments looks promising. There are numerous ways in which AI could help financial institutions enhance their payments operational efficiency, AML and fraud prevention, as well as customer satisfaction. Nevertheless, being able to deliver that promise requires doing it well. AI technologies are nascent and evolving quickly – and therefore implementing AI in any payment use case presents unique risks and challenges to financial institutions. These include data quality, ethical and legal concerns, computational resources, interpretability and explainability, and above all security.

To avoid unintended consequences and minimize risk, FIs must ensure they choose the right partner who will enable them to harness the power of AI technologies responsibly. At Finastra we continuously explore, experience, and adopt advanced technologies alongside our partner ecosystem solutions. As such we empower our customers to harness the power of AI safely and responsibly in order to boost their business performance and customer satisfaction. We have therefore established AI usage policies, processes, and best practices across all functions in our organization.

Contact us to learn more.