How UT Austin Built an AI Tutor Platform, and What ET Is Doing to Scale It

When instructors at UT Austin want to give students a study partner that stays close to course material, avoids simply handing over answers, and remains available whenever students need help, they are increasingly looking to SAGE.

SAGE is the university's AI-powered tutoring platform. It allows faculty to create course-specific Socratic chatbots that work inside Canvas, giving enrolled students access through the same course navigation they already use every day.

The platform is developed and operated by the Office of Academic Technology under the Office of the Provost, with Enterprise Technology serving as a key operational partner. That partnership gives SAGE something many promising academic tools never reach: a path from small pilot to durable institutional service.

What SAGE is built to do

SAGE is designed around a clear instructional philosophy. The goal is not to provide students with fast answers. The goal is to help them think. Faculty upload course materials, define the boundaries of what the tutor can discuss, and shape the tutor's behavior around course objectives and likely student misconceptions.

The resulting chatbot uses a Socratic approach, prompting students to reason through questions rather than retrieving solutions outright. Conversations stay intentionally constrained to course-relevant content, which makes the platform feel less like a general-purpose chatbot and more like a teaching tool.

A product model shared between OAT and ET

The partnership between the Office of Academic Technology and Enterprise Technology reflects a model UT Austin has been refining across emerging digital services: academic technology teams stay close to product direction and faculty needs, while central IT helps provide the operational backbone required for reliability, security, and scale.

For SAGE, that means OAT owns the product roadmap, vendor relationship with Anthropic, governance and compliance framework, and the faculty-facing support model. Enterprise Technology contributes the platform operations layer, including AWS cloud infrastructure, Canvas integration support, platform engineering, and Tier 1 technical support through the ET Service Desk.

That structure matters because it keeps the service close to teaching needs without sacrificing the expectations that come with an enterprise platform. SAGE is able to remain faculty-centered while still operating inside a more durable institutional support model.

Built for trust, not just speed

SAGE's path to adoption has involved more than technical integration. The platform has gone through review by UT Austin's Information Security Office, Digital Accessibility Office, Office of Legal Affairs, and the university's Fair Use Librarian, reflecting how seriously the team has taken trust, compliance, and instructional responsibility.

Faculty governance bodies were also brought in early. The platform received formal briefings before faculty-facing groups and included public comment and working-group input that pulled in faculty, students, and ethical AI perspectives. That review process helps explain why SAGE is being scaled deliberately rather than rushed.

• Student interactions are intended to stay within UT's environment rather than feeding third-party AI training or marketing pipelines.
• Faculty remain in control of whether and how SAGE is used in their courses.
• The platform operates within UT Austin's broader responsible AI framework for teaching and learning.

What adoption already shows

The current phase is still a structured pilot, but existing ET reporting already shows meaningful adoption signals. In the January 2026 CIO Updates published in this site, ET noted that in Fall 2025 more than 2,100 students used 366 tutors across 90 Canvas courses, generating 128,000 interactions.

That earlier scale matters because it shows SAGE is not beginning from zero. It is moving into its next phase with evidence of real student use, real faculty experimentation, and a growing need for stronger support structures around the platform.

Your draft also points to a Spring 2026 view of the pilot with 34 actively engaged faculty and more than 900 students interacting with course-specific tutors. Taken together, those signals suggest not just curiosity, but sustained course-level integration among faculty who are using the platform intentionally.

What comes next

Fall 2026 is shaping up as the next major transition point for SAGE. The target described in the draft is 100 to 150 actively engaged faculty, with expansion paced around support capacity rather than raw headcount. That is a healthy sign. It suggests the team is treating adoption quality as seriously as platform availability.

A key part of that next phase is the addition of a SAGE Program Manager role in the Office of Academic Technology. The position is intended to strengthen faculty engagement, coordinate the SAGE Faculty Fellows program, build training pathways and exemplars, and serve as connective tissue across OAT, ET, and college-level support units.

What this means for faculty

For faculty, SAGE represents a concrete way to experiment with AI in teaching without starting from scratch. The product is already integrated into Canvas, shaped by instructional design principles, and backed by a support model that splits responsibilities in a practical way: ET handles front-line technical issues, while OAT stays focused on product evolution and faculty success.

Faculty interested in learning more can visit sage.utexas.edu. Questions about platform use and academic support can go to the Office of Academic Technology, while the ET Service Desk remains the right first stop for authentication, access, and general technical troubleshooting.