The Future of AI in Powering Payments

Intro

Growing transaction volume and demand for fraud prevention has caused many financial institutions (FIs) to look towards artificial intelligence (AI) to more effectively serve cardholders. AI is used to provide more seamless financial journeys through increased personalization and enhanced customer service, and to support FIs in their back-end security and payment processing functions.

Cardholder journeys are enhanced through predictive analytics, which help offer insights into cardholder spend and assist in predicting future spending behavior, as well as through natural language processing, which powers customer-facing chatbots. To mitigate risk and ensure regulatory compliance, machine learning is used to improve fraud prevention by identifying suspicious patterns and flagging potentially fraudulent accounts and transactions, while generative AI helps generate payment processing documentation.¹

Future of AI in Payments

By analyzing vast amounts of transaction data in real-time, FIs can use AI to identify patterns and anomalies that may indicate fraudulent activity as well as confirm valid transactions. This reduces instances of false transaction declines, a major pain point for cardholders (nearly half of consumers won’t retry a payment following a false decline²). AI further improves FI’s fraud management through upgraded fraud scoring and assistance with investigations. By reducing instances of fraud, AI can improve the quality and accuracy of data flowing through payment networks.

When determining risk for new cardholders, FIs benefit from AI’s ability to analyze a broader set of data points. In the credit decisioning and underwriting process, this allows FI’s to more accurately predict spend behavior and loan repayment. By assessing credit worthiness based on factors beyond a potential borrower's credit score, FIs can deliver more refined risk assessments and extend credit to a wider customer base, including traditionally underserved segments.

Risk management processes continue to be enhanced by AI’s ability to prevent failures in backend business operations. For example, TransFund owned ATMs use machine learning technology to analyze service performance and trends, providing preventative and predictive actions while optimizing real time field service performance. FI’s compliance and regulatory processes can also benefit from AI integration as the technology can be used to review legal and regulatory documents and extract key data points and clauses. AI also assists in code generation for tasks such as automating the processing of complex regulatory documents. This can significantly reduce the time and resources required for document review processes.

From the cardholder’s perspective, AI helps address a growing demand for increased personalization with financial services. FIs can use next best action recommendations generated by AI to determine when to contact a particular cardholder, as well as which channel to use and what to say. To further drive engagement, FIs can leverage AI to provide personalized insights, reminders, and recommendations, such as flagging higher-than-usual automatic billing transactions or reminding account holders to transfer funds and avoid overdraft fees.

The cardholder experience continues to be enhanced with AI-supported customer servicing. Chatbots and Al-powered virtual assistants can be used to respond to cardholder queries and enable self-service capabilities, saving FI’s time and money that can be spent on higher value cardholder interactions, such as cross-selling new products. FIs can also use AI to search for, collect, and analyze feedback to improve their business operations and provide cardholders with a more seamless financial journey.

Considerations

While the use cases and benefits of AI continue to grow, FIs must be aware of the risks introduced by this new technology. In the same way AI provides FIs with more tools to operate efficiently, it also provides fraudsters with more tools to beat standard security measures. Bad actors have already begun using this technology to generate fake identities, backed up with fake social media accounts and fake ID documents.³ Additional issues arise when FIs are unable to differentiate between their cardholders and AI. For example, when cardholders use bank aggregator software to receive financial insights across multiple financial accounts, it is difficult for FIs to determine when the cardholder is logging into their account vs. when an AI bot is logging into the account. If it is the AI bot, it can be unclear if the cardholder has given it the authority to do so.³

FIs also face obstacles of integration when attempting to combine this new technology with legacy infrastructure. For AI to be successfully integrated, FIs must take the time to understand what backend operations and frontend services would benefit from AI integration, and to what extent AI will take over existing processes. Additionally, while AI models are built with defined parameters and training data, deep learning models develop internal logic that is not easily interpretable. This ‘black box’ nature can make it difficult for FIs to fully understand how a decision (e.g., a credit line approval) is reached.

Conclusion

As the capabilities and applications of AI continue to grow, the payments industry has become an active space to showcase its many use cases. It has the capability to improve the cardholder experience through personalized messaging, insights, and servicing, and to support FIs with fraud mitigation and regulatory compliance. That said, AI’s advancements in the payments industry also carry risks, most notably misuse from bad actors looking to profit at the expense of FIs and their cardholders. As AI becomes a permanent fixture, FIs will need to assess their AI readiness, determine the manner and the extent to which AI will shape their business, and develop a strategic roadmap for integration.

Sources:

• Mastercard Expertise
• Checkout.com, How is AI affecting the payments industry?, 2024
• Wired, How artificial intelligence is transforming payments