TD Bank is prioritizing AI and generative AI in internal operations and aims to study consumer needs before deploying gen AI for client facing operations. 

“We’re currently focused on leveraging gen AI for internal operations, mainly employee-facing, human-in-the-loop experiences that lead to better efficiency, better productivity — and all of that ends up benefiting our clients,” Ted Paris, executive vice president and head of analytics, intelligence and AI at TD Bank, told Bank Automation News.

Streamlining backend operations with AI and gen AI allows the $1.4 trillion bank to better serve consumers at a lower cost, Paris said.  

The bank is taking a slow approach to providing gen AI capabilities to its chatbot TD Assist, Paris said. TD must understand consumer needs before deploying new tech, and any new capabilities must have a “meaningful playing impact on consumers,” he said. 

“What we do ultimately is with a customer outcome in mind,” Paris said. “Deploying consumer-facing AI might be something we consider at some time in the future and when the time is right. But right now, we’re focused on leveraging gen AI internally.” 

Slow gen AI approach

Banks in general are treading cautiously when it comes to consumer-facing gen AI, Dimitrios Papanastasiou, global head of gen AI solutions at Moody’s Analytics, previously told BAN.  

“Hallucinations and uncertain regulations are the biggest holdbacks,” he said.

Banks are first deploying gen AI internally, where they find more cost savings, Papanastasiou said, adding that it will take time for FIs to find ways to monetize consumer-facing gen AI. 

Internal uses at TD 

TD has been using AI and machine learning for years for cybersecurity, anti-money laundering and fraud detection, Paris said.  

“We’re looking at utilization of gen AI as highly focused around our inwardly pointing tasks,” he noted. 

Gen AI currently in use at TD include: 

• KYC; 

• Coding assistance; 

• Internal research chatbots; and 

• Customer representative assistance. 

“I won’t cite any particular metrics at this stage around the gen AI piece … but we have found great success in deployed areas,” Paris said. 

The Toronto-based bank, which partners with its cloud provider Google Cloud to develop data-driven AI solutions, according to an April release, also relies on its in-house AI developer, Layer 6, which it acquired in 2018 for an undisclosed amount. Layer 6 provides most of the bank’s AI innovation and development, Paris said, noting that TD works with vendors when a solution needs to be scaled.