TD Bank has deployed AI-driven platform TD AI Prism to predict consumer needs and provide them with solutions proactively. 

The $1.2 trillion bank’s marketing teams use the platform to predict what retail products and services clients might need so it can target communications for them, such as identifying consumers looking for a mortgage, Sumee Seetharaman, head of AI and machine learning, told FinAi News.

“Traditional predictive AI models can only typically perform a single task. We’ve had success with them in the past, but they’re inefficient,” Seetharaman said. “Each time we developed a new use case, we had to rebuild our models.” 

TD AI Prism synthesizes significantly more data points to perform multiple tasks and predict multiple events, allowing the bank to build more use cases on top of this foundational model, she said. 

“We’ve found that TD AI Prism is 20% to 30% better at identifying our clients’ needs compared to our other existing models,” Seetharaman said. 

The Toronto-based bank built TD AI Prism in-house with its Layer6 team after three years of research, he said. The bank didn’t disclose the amount that was spent to develop the platform. 

“It’s built on transformer architecture and combines a large-scale AI model with bank datasets to better understand our clients,” Seetharaman said. 

The platform was launched in beta in June, and the bank is monitoring “its performance over the coming months before we roll it out further across the bank,” she said. 

Partnership with MIT 

TD teamed up with educational institutions like MIT to research additional use cases for gen AI and agentic AI in banking and financial services, Seetharaman said. 

“As part of our involvement in the MIT sAIpien program, we’re participating in joint research on AI alongside a cohort of collaborators across key sectors,” she said. “The group is interested in tackling moonshot ideas like reimagining the future of banking with AI.” 

Working with MIT, TD plans to develop gen AI-driven technologies that will amplify human efforts rather than replacing them, Seetharaman said. 

Register here for early-bird pricing for the inaugural FinAi Banking Summit 2026, taking place March 2-3 in Denver. View the full event agenda here.