Blogs

Member Spotlight: Q&A with Zain Kazmi, Associate Vice Chancellor, Chief Digital & Analytics Officer at University of Texas System and Peter McCaffrey, pathologist and Chief AI Officer at the University of Texas Medical Branch.

Q: Tell us about the University of Texas (UT) system and your roles within the organization.

Zain Kazmi: The UT system consists of 13 institutions, eight of which are focused on healthcare or have a healthcare presence. Each campus operates in different markets across Texas. My role as Associate Vice Chancellor covers two main areas: strategy—which includes the One UT Health Strategic Framework and 27 collaboratives spanning our health enterprise—and digital/tech, which includes overseeing data and AI initiatives and partnerships with major tech companies.

On the education side, UT’s footprint is significant: we have seven medical schools, ten nursing schools, three pharmacy schools, two dental schools, and over 30 different healthcare programs. This makes our educational and research presence one of the largest in the country.

Peter McCaffrey: I’m a pathologist and Chief AI Officer at the University of Texas Medical Branch (UTMB), one of the eight health institutions at UT. I oversee AI efforts across both the academic and health system sides, and I chair the UT REAL Health AI program under the UT Health Intelligence Platform (HIP).

Q: What is UT HIP, and what does it aim to achieve?

Zain Kazmi: UT HIP is the University of Texas Health Intelligence Platform. Launched in 2016, UT HIP was created to unify the data assets of the eight healthcare campuses, particularly by centralizing and standardizing data sharing. Before UT HIP, there was no mechanism to bring together data across the system into one place, which limited collaboration and improvement. HIP is funded by both the health institutions and the UT system and is overseen by a multidisciplinary council, of which I am a program sponsor. Its mission is broad, serving quality officers, business development, employee health, research, and more.

Peter McCaffrey: As chair of the AI component of UT HIP, I get to lead a federated, deliberate effort to align methodologies and infrastructure across the system. In AI, we build a shared roadmap, synchronize priorities, and develop common approaches for assurance, validation, and collaboration. UT HIP helps us move forward together as a system and mature appropriately through shared experience and knowledge.

Q: What have been the key successes and learnings from UT HIP over the past decade?

Zain Kazmi: A lot has happened in 10 years since launching HIP. We’ve built a system-wide culture where data is used transparently to identify strengths and weaknesses, especially in quality improvement. Before HIP, many leaders across campuses did not know each other. Now, they meet regularly to review and act on data together. This collaboration to improve quality is central to our mission at UT.

And, with hundreds of thousands of lives covered under UT’s health plan, UT HIP has enabled deep dives into medical and pharmacy spend, improving both care quality and cost management for our employees and the system.

Q: Why did UT decide to join CHAI, and what were you hoping to achieve?

Zain Kazmi: UT wants to lead in AI, not just in Texas but nationally and globally, in outcomes, policy, and regulation. Joining CHAI, the leader in health AI, as a founding member was important to us. Joining CHAI aligns with our recent launch of a focused AI program under UT-HIP, helping us continue to set the pace in this rapidly evolving field.

Peter McCaffrey: At our scale in Texas, it’s crucial to align and streamline how we approach AI challenges across our campuses and organizations throughout the state. If each campus or organization does things differently, progress is limited. Joining CHAI helps the UT System establish frameworks and best practices so we can move forward together as a unified health ecosystem and continuously drive growth and improvement.

Q: How is UT managing AI in research, education, and governance?

Zain Kazmi: Governance is critical and evolves constantly as leadership and needs change and as technology progresses. At the highest level, a multidisciplinary committee—one leader from each campus—makes major decisions about scope and growth trajectory. For research, we have a dedicated committee that addresses how data assets are used and navigates processes related to IRBs, grants, and controls. There are also subcommittees for architecture, employee health, and quality, all reporting up to the Office of Health Affairs.

Q: What about the procurement of AI tools—how does that process work at UT?

Zain Kazmi: Procurement is matrixed. Some tools are acquired through long-standing partners, while others involve more system-level coordination. We’re continually looking for ways to improve this process, especially as AI becomes more central to our operations.

A significant driver of value is the unified approach across UT campuses. We're maturing our AI procurement processes by openly sharing experiences with products, models, and tools, fostering dialogue about lessons learned and best practices for improvement.

Q: How do you envision CHAI’s resources and best practices helping in clinician education and training?

Peter McCaffrey: The big narrative here that's coming up everywhere in our system is about upskilling our workforce to become proficient users of AI, whether that’s for care delivery or administrative tasks. This can be a daunting task if everyone is working off different definitions or going at it in separate ways. What’s powerful about CHAI’s best practices and playbooks is that they enable us to educate clinicians in a consistent way—using common frameworks, best practices, and agreed upon standards and definitions. We’ve been able to implement this in other areas of medicine, and AI shouldn't be any different. The more we can align on these specifications, the better it will be for clinicians, patients, and our entire system.

Q: Where do you hope to see the most progress for AI in health?

Zain Kazmi: One major area I’m hopeful for is the workforce. In Texas, we’re facing access issues—in large part due to not having enough clinicians, even with our existing pipeline of medical and nursing schools. The answer isn’t just more schools; it’s also about technology. AI must be central to helping our workforce reimagine care delivery. Technologies like agentic AI, productivity tools, ambient listening, and clinical dictation services can all contribute to solving these challenges. Looking five to ten years ahead, leveraging AI will be essential to make our workforce more productive and better able to serve our patients in the state of Texas and beyond.

Peter McCaffrey: I fully agree. The situation is urgent. The healthcare people want is care with immediate access to expertise, anytime, anywhere. That can’t happen at scale without an augmented workforce supported by AI. I see AI being especially promising for information retrieval and summarization, non-clinical task automation for things like early patient intake and post-discharge checkups, and workforce upskilling to help clinicians become proficient users of AI and provide the highest quality healthcare to Texans.

Q: What are you most excited about UT joining CHAI?

Peter McCaffrey: We are excited about CHAIs model card specifications and assurance resources– two really important elements that lead to a win for everyone involved. At UT, we want to run at something, we just need a direction to run in, and CHAI provides that direction.

Zain Kazmi: I am most excited about becoming part of a community of like-minded leaders who share our mission and are committed to shaping the future of AI in healthcare. CHAI brings together experts across institutions who are willing to compare notes, collaborate, and approach challenges with shared purpose—something we place high value on at UT.