AI Applications Expand for Merchants Despite Potential Hazards
In the rapidly evolving world of e-commerce, the integration of Agentic Artificial Intelligence (AI) is revolutionising merchant operations and fraud detection. This advanced form of AI is taking centre stage, not just as an assistant, but as an autonomous decision-maker, driving efficiency, customer service, inventory management, dynamic pricing, and fraud prevention.
According to a report by Don Apgar, Director of Merchant Payments at Javelin Strategy & Research, the applications of AI in the payments ecosystem cover a broad spectrum, from transaction routing to regulatory compliance. The role of AI in Anti-Money-Laundering (AML) monitoring and transaction routing is expected to remain significant, regardless of regulatory changes in the current administration.
Currently, Agentic AI is making waves in merchant operations and e-commerce through autonomous shopping agents, inventory and supply chain optimisation, dynamic pricing and operational decisions, and customer service automation. For instance, AI agents independently research, compare, and purchase products based on learned preferences and goals, streamlining merchant workflows and customer interactions.
Looking ahead, Agentic AI is expected to take greater control in managing complex merchant tasks and fraud mitigation, with minimal human oversight. The potential future applications include proactive fraud detection and prevention, adaptive learning against fraud tactics, and integrated fraud and merchant operations.
However, it's crucial to note that AI should not be in the critical path of any workflow, as it can hallucinate (make up information that's not there). A buffer around any public-facing AI initiatives is necessary to mitigate potential mistakes. Fine-tuning is also essential in how AI models analyse their findings and present conclusions to avoid unintended consequences in fraud response.
AI is not yet capable of handling 100% of inquiries or tasks, and some consumers express reservations about AI making purchases on their behalf due to potential mistakes, overspending, or data disclosure. Nevertheless, the efficiencies of AI should be leveraged, but it should never create a point of failure in a workflow. A human backup, or a backstop, should always be in place.
Major players in the payment processing industry, such as Visa and Mastercard, have already rolled out platforms built to harness Agentic AI. AI is being implemented in customer-facing situations, particularly in chat use cases, allowing for minimal human interaction in shopping and making purchases.
Examples of AI implementation in e-commerce can be found in the apps of major retailers like Walmart (Sparky) and Amazon (Rufus). AI is frequently implemented in fraud detection, particularly in card-not-present environments like e-commerce, where it can identify patterns and red flags.
In conclusion, Agentic AI is evolving from assisting humans to independently managing commerce processes. Its role in fraud detection is poised to become more autonomous and dynamically adaptive, enhancing security and operational resilience in the e-commerce ecosystem. While there are challenges to be addressed, the potential benefits of Agentic AI in streamlining commerce and enhancing fraud detection are undeniable.
Businesses in the e-commerce sector are increasingly leveraging artificial-intelligence (AI) technology, not only as assistants but also as autonomous decision-makers, thanks to Agentic AI. This advanced form of AI is revolutionising fraud detection and is expected to take control in managing complex merchant tasks, including proactive fraud detection and prevention, adaptive learning against fraud tactics, and integrated fraud and merchant operations (finance). Moreover, major players in the payment processing industry, like Visa and Mastercard, have already adopted AI technologies in customer-facing situations, particularly in chat use cases and fraud detection, especially in card-not-present environments (technology).
