TD Bank Group Launches GenAI Pilots to Enhance Customer Experience

The Toronto Dominion Bank Group (TD) announced two generative AI pilots to support streamlining of customer experiences in its contact centers and accelerate the coding flow for its engineers.

Both the contact center and the engineering pilots have been developed through TD Invent, the Bank's enterprise approach to innovation. The approach enables TD to assess how emerging technologies, such as generative AI and spatial computing, can solve real-world problems.

Luke Gee, Head of AI and Analytics at TD, says, “Generative AI is transformative; its potential to revolutionize how we deliver for our colleagues and customers is incredible.” Gee adds that the early exploration of generative AI is an evolution of the organization’s ongoing AI strategy.

As announced, TD has developed a generative AI virtual assistant that will help contact center colleagues retrieve answers to regular banking inquiries in seconds. Layer 6, TD’s development and research center, has helped develop the generative AI virtual assistant and train it alongside colleagues with a deep understanding of TD policies and procedures.

The generative AI assistant sources its answers from TD policies and procedures and provides summarized responses utilizing conversational language to contact center colleagues. This allows colleagues to answer customer inquiries faster. The bank will test the generative AI virtual assistant with front-line contact center colleagues this year.

In addition, TD has launched a pilot for GitHub Copilot, a generative AI programming assistant developed by Microsoft, to augment its engineering function. At TD, the GitHub Copilot assists engineers by analyzing when they write code and providing real-time suggestions for completing and testing it.

Further, the platform enables a more efficient developer experience by helping engineers develop maintainable code.

As per a TD report, the bank conducted a pilot program with a selected group of engineers between September and December 2023 to evaluate GitHub Copilot for developing unit testing code. After surveying the participants, TD discovered the following:

• 50% of the engineers reported that GitHub Copilot could save them up to 20 hours during a two-week sprint.

• 93% of data engineers indicated that GitHub Copilot made them equally or more productive.

• 75% of the engineers found it easy to incorporate GitHub Copilot into their daily tasks.

"We expect that generative AI is going to play an important role in how we evolve our engineering function and culture at TD," says Licenia Rojas, Senior Vice President, Chief Engineer, and Architect at TD.

"By implementing solutions such as GitHub Copilot, we're helping empower our talented engineers to focus on new opportunities and demonstrating our commitment to delivering a positive engineering experience and to developing the best technology talent in the industry," Rojas adds.