University of Michigan Launches Agentic AI Virtual Teaching Assistant

At the University of Michigan's Stephen M. Ross School of Business, a new Virtual Teaching Assistant pilot program is utilizing agentic AI to provide students with 24/7 access to support and self-directed learning. Powered by Google's Gemini models, the Virtual TA can be customized for specific curricula and is designed to "foster critical thinking by never giving away answers, ensuring students actively work towards solutions," according to a news announcement.

The Virtual TA acts as a practice partner, guiding students through problem-solving and providing on-demand explanations of complex course concepts. The tool also provides real-time analytics on student engagement, summaries of commonly asked questions, and feedback on the AI agent's effectiveness. It is being piloted in courses such as financial technology, operations strategy, operations analytics, and statistics, where early results "indicate greater student engagement and deeper insights into the learning patterns of students for instructors," the announcement said.

"By partnering with Google Public Sector, we're integrating today's leading technology to improve the educational experience for our students," said Jun Li, Michigan Ross Professor of Technology and Operations, in a statement. "We've seen early indications of success in leveraging the AI-powered Virtual TA and look forward to further using this tool to advance learning."

"Google Public Sector is proud to pioneer this personalized learning approach to improve educational delivery, provide early identification for intervention, and help educators solve some of their most pressing challenges," commented Chris Hein, field chief technology officer for Google Public Sector. "The Virtual TA is a prime example of how a customizable, Gemini-powered solution can advance the future of learning and help train the next generation workforce."

The university plans to incorporate insights from its Virtual TA pilot into a larger research study across multiple higher education institutions, involving 9,000-plus students, 72 courses, and 26 schools. The goal: to develop guidelines on deploying the technology in diverse educational settings and better understand how colleges and universities can adopt AI for learning.

About the Author

Rhea Kelly is editor in chief for Campus Technology, THE Journal, and Spaces4Learning. She can be reached at [email protected].

Featured

  • networked node grid featuring illuminated connectors

    Infrastructure Accounts for More than Half of Worldwide AI Spending

    Gartner forecasts worldwide AI spending will reach $2.67 trillion in 2026, up 49.5% from 2025, with AI infrastructure accounting for nearly $1.5 trillion, or about 56% of the total. For every dollar Gartner expects to be spent on generative AI models, more than $52 will be spent on infrastructure.

  • CIS Sandbox Tutor Eli Blouin at Bentley University imagines a future technology learning partner that supports curiosity, critical thinking, feedback, and personalized learning.

    Student Voices: Technology as a Future Learning Partner

    A data analytics and marketing major at Bentley University and a distinguished lecturer at the university explore the future of "technology learning partners" — technologies that participate in the learning process by asking questions and giving feedback, not just information search results.

  • young man hand working on a laptop, bokeh blurred at night

    Redesigned Copilot Will Build Apps and Work on Its Own

    Microsoft recently announced that it's redesigning Copilot, the company's all-encompassing AI assistant, so users can hand off complex tasks, build tools by describing what they need and assign work to an agent that keeps going after they log off.

  • Abstract glowing network nodes forming conversation bubbles

    OpenAI Introduces Dots, AI Agents that Continue Working Between User Conversations

    OpenAI has announced Dots, GPT-6 Astra-powered agents that run on their own cloud computers, using connected applications to carry on with assignments between conversations and incorporating users' feedback as they go.