University of Chicago accelerates secure generative AI adoption with Royal Cyber, powered by Microsoft Azure

The University of Chicago (UChicago) is one of the nation’s most prestigious research institutions, serving more than 35,000 students, faculty, and staff. As interest in generative AI surged across campus, the university faced a challenge familiar to many: third-party tools were being used without institutional oversight, creating security, privacy, and governance risks.

To address this, the university partnered with Royal Cyber—recognized for secure, enterprise-grade AI platforms for higher education. Working closely with Microsoft and UChicago’s IT Services team, Royal Cyber helped design and deploy PhoenixAI—a secure, scalable generative AI assistant powered entirely by Azure.

Advancing progress without slowing momentum

Faculty and students wanted access to generative AI. But without a centralized solution, the university risked both fragmented adoption and shadow IT usage. Unapproved tools lacked basic data protection and compliance controls, which put institutional data at risk.

Royal Cyber worked with the university to co-develop a framework that met security and governance standards while preserving flexibility for academic exploration. The goal: deliver an institutionally managed AI solution built on trusted Azure infrastructure, deployable before the semester’s start, customizable by department or course, and capable of serving tens of thousands of users without adding operational overhead.

PhoenixAI was built entirely on Azure using a modular, container-based architecture designed for performance, observability, and scalability. The system architecture was deployed fully within UChicago’s Azure tenant, ensuring that the university retained ownership of its data, infrastructure, and access policies.

PhoenixAI integrates a wide array of Azure technologies to deliver secure, scalable performance, including:

Fast delivery through collaboration

The PhoenixAI project launched on a tight timeline—just six weeks before the start of the academic semester. Royal Cyber adopted an agile delivery model, beginning with stakeholder workshops with IT Services, faculty, and compliance officers. Requirements were quickly documented and translated into a scalable and compliant Azure-based architecture.

Model integration began with GPT-4o and later expanded to include GPT-4.1, GPT-5, and GPT-o3, giving users access to advanced reasoning capabilities. Royal Cyber embedded custom throttling and role-based access modules that allowed IT administrators to define usage limits per user group, course, or department. Faculty members could set permissions at the course level, ensuring appropriate use of generative AI tools within academic boundaries. PhoenixAI saw four major updates within six months, proving its agility and responsiveness to user needs.

Royal Cyber also provided expertise in deploying modular agents via Azure AI Foundry. Faculty and researchers were able to build and deploy GPT-powered agents tailored to administrative tasks, academic research, or student support—all governed by university policy and deployed securely within PhoenixAI.

Broad adoption, high trust, and significant cost savings

Within six months of launch, PhoenixAI reached more than 35,000 users, with 25% active engagement. Compared to commercial SaaS alternatives, AI usage costs dropped by 95% (based on internal analysis).

Faculty members use PhoenixAI to build agents for administrative workflows and research. Students rely on it for tutoring, resume writing, and academic support. Researchers request dedicated API access for experiments.

For UChicago IT Services, PhoenixAI provided centralized oversight and audit-readiness through integrated logging, consumption monitoring, and policy enforcement. Real-time observability ensures that security and performance standards are continuously met. The platform also integrates with external sources like Google Drive, OneDrive, and Box, empowering users to securely access and interact with content across institutional and personal storage systems.

Today, PhoenixAI continues to evolve. Future plans include expanding access to clinical departments, integrating with more academic systems, and developing agentic features that allow the AI to take proactive actions based on user context.

The university also plans to enhance its research capabilities by building APIs and tools that support advanced experimentation across disciplines.