TD Bank and the MIT Media Lab convene students, researchers, and bank leaders to reimagine the branch
TD's membership in the MIT Media Lab community reflects a shared understanding: that the digital future should also be a human one. The Bank's "Digital First. Human Always" principle and its broader purpose — to reimagine what banking can be for clients, colleagues, and communities — find a natural counterpart in the Media Lab's mission to invent a better, more just future, by, with, and for all. Human-centered design is a guiding principle across TD's approach to the branch, with each decision tested against a single question: how does it affect the client experience? That intent is most visible through Layer 6, the AI Center of Excellence within TD's broader AI2 group, whose research informs how the bank serves more than 10 million clients.
“Spending time with the teams at the MIT Media Lab was a reminder that great ideas come from a strong sense of curiosity and innovation.
What stood out was how consistently the thinking came back to the same goal – using technology to make the experience simpler and more personal for clients,
said Sumee Seetharaman, Vice-President of AI and Machine Learning Practice at TD.
One day, one central question, five distinct concepts
The collaboration took form in May, when about 35 Media Lab researchers and MIT Sloan School of Management students gathered with a cross-functional TD cohort for the TD x MIT Media Lab Hackathon: The Future of Banking/The Future of the Branch. TD has been studying the future of the branch for some time, including human-centered "living experiment" stores in New Jersey, Florida, Massachusetts, and Ontario—work that has even extended to rethinking the "ATM" as an Automated Treat Machine for dogs.
The brief asked participants to reconsider the branch as an experience layer in which physical presence, digital intelligence, and human judgment intersect. Five teams presented concepts evaluated against four criteria: innovation, impact and scalability, readiness and feasibility, and ethical considerations. The submissions included an AI co-pilot to prepare small business owners for more substantive advisor conversations; a tiered branch network organized around the principle "App for Transactions. Branch for Decisions."; a regional economic risk model built on approximately 150 million coordinated AI agents; a shared intelligence system intended to make each branch visit better prepared and more personalized; and a smart-glasses concept that surfaces prompts when day-to-day spending diverges from goals established in-branch.
Himanshu Rai, TD’s Lead Scientist for the US Layer 6 team, served as a judge for the hackathon. “I was impressed by the thinking and insights shared by the teams who participated,” he said.
There was a great variety of ideas, but they converged on a similar vision of the future: how AI acts as a layer that augments human judgment rather than replacing it.