Authored by stclarke, this article explores UNSW Sydney’s integration of Scout, an AI agent developed with Microsoft Azure OpenAI and Copilot Studio, to transform student support and administrative efficiency.

UNSW Simplifies Student Life with Agentic AI: Scout Powered by Microsoft Azure OpenAI and Copilot Studio

Author: stclarke

Introduction

The University of New South Wales (UNSW) in Sydney, Australia—serving over 70,000 students—continues its mission to enhance lives through innovative research, transformative education, and social justice. Recognizing the pressures faced by both staff and students, UNSW’s Educational Technology Support team focused on reducing administrative burdens and leveraging new technologies to streamline support, thereby improving teaching, learning, and community engagement.

The Role of the Educational Technology Team

Managed by Pradhiban Duraisamy, the Educational Technology Support team comprises a group of “education technology enthusiasts” who have for years been dedicated to implementing solutions that:

  • Lower administrative workloads
  • Improve digital accessibility
  • Enhance student and staff wellbeing

Introducing Scout: UNSW’s Agentic AI Solution

Aligning with the university’s Educational Technology Roadmap 2024-2028, the team piloted Scout, an AI-driven agent powered by Microsoft Azure OpenAI and built using Microsoft Copilot Studio and Microsoft Power Platform.

Scout, distinguished by its yellow hoodie (reflecting UNSW’s colors) and welcoming demeanor, serves 7,500 pilot students. Its core objective is to quickly and reliably answer common queries, thereby making access to essential information easier, building student confidence, and freeing educators for deeper, more meaningful engagement.

Democratizing Information Access

Navigating thousands of web pages for course details, schedules, and society information can overwhelm students, especially international students unfamiliar with institutional terminology and acronyms.

Through consultation with teaching staff, the technology team learned that students routinely ask about:

  • Assessment dates
  • Enrollment status
  • Available support services

Scout addresses these repetitive queries, providing a straightforward, reliable way to source answers—students need not navigate ownership of various online resources.

Located within the learning management system (Moodle), Scout draws information from multiple sources in a hierarchical structure:

  1. Course-specific data: Custom class guidelines and processes, gathered via Microsoft Power Apps.
  2. Public course outlines: Contacts, learning outcomes, assessment guides, and lecture slides.
  3. UNSW websites: Policies, program details, clubs, societies, and services.
  4. Azure OpenAI GPT: Used when no adequate answer is available from other sources.

Scout’s responses include links to original UNSW sources for verification.

Early Results and Expansion

The pilot expanded from 1,000 to 7,500 students due to positive results. Notable metrics include:

  • 58% of Scout usage occurs outside normal hours, bridging support gaps and reducing staff pressure
  • Routine testing has shown no issues with accuracy among surveyed students
  • Students are enthusiastic about expanded capabilities, such as:
    • Course recommendations tailored to career goals
    • Suggestions for prerequisite or complementary subjects

Benefits to Lecturers and Academics

Scout’s automation of routine administrative questions has shifted student engagement with lecturers towards deeper academic topics. As a result, educators can focus on their expertise rather than logistical queries, aligning with the university’s vision to prioritize academic interaction.

Promoting Student Wellbeing

Duraisamy outlines how Scout may further improve student wellbeing:

  • Equitable learning support with appointment booking and memory of conversations
  • Empathetic responses to requests for learning adjustments or additional support
  • Integration with health and psychological services, especially during stressful periods (e.g., exams)

Broader Impact and Future Directions

Duraisamy anticipates further adoption of AI agents at UNSW, with specialized teams using Microsoft Power Platform’s low-code tools to manage their own information agents. This decentralizes knowledge management away from central teams, enhancing flexibility and responsiveness.

Advice for Other Institutions

Duraisamy encourages other educational institutions to:

  • Keep development in-house wherever possible for agility and deeper learning
  • Pilot solutions over discrete periods, iterating based on real-time feedback from users

He believes starting such projects is both feasible and rewarding, opening up diverse applications of new AI technologies in education.

Conclusion

UNSW’s Scout exemplifies how agentic AI—built with Microsoft Azure OpenAI and Copilot Studio—can transform student support, improve administrative efficiency, and support both wellbeing and academic engagement. Its success positions UNSW as a model for other education providers considering AI-powered solutions.

This post appeared first on “Microsoft News”. Read the entire article here