On the 63rd episode of Enterprise AI Innovators, hosts Evan Reiser (co-founder and CEO, Abnormal AI) and Saam Motamedi (General Partner, Greylock Partners) talk with Lev Gonick, CIO at Arizona State University, about what it takes to move past pilots and into durable systems. The theme that comes up repeatedly is agency, for students, for staff, and for IT itself. “We want our students to succeed,” Lev says, and in the AI era he wants students to “have agency, the autonomy to actually participate themselves” in how AI gets used in their education and problem-solving. That framing matters because it pushes the conversation away from policy-only debates and toward concrete decisions: tools, workflows, guardrails, and where the institution chooses to invest.
ASU does not talk about AI like a feature. Lev talks about it like an operating system for a university that runs at the scale of a global enterprise. ASU serves almost 200,000 students, spans campuses in Tempe, Los Angeles, and Washington, DC, and supports 35 campuses worldwide, plus education operations for the U.S. Air Force and the Army. For a CIO, that footprint forces a practical question: if you measure your institution by student success, what do you build when AI becomes a default capability?
A big part of ASU’s approach is treating AI as both research infrastructure and enterprise infrastructure. Lev describes a research environment with roughly 6,000 faculty and about $350 million in sponsored research tied to AI-related work, spanning domains like energy, water, and core machine learning models. Supporting that range is not just about model access. It requires platforms that can handle security and compliance, and maintain a steady cadence of experimentation without turning every request into a bespoke IT ticket.
So ASU built internal capacity early. He formed an AI acceleration team within enterprise IT, starting with about 20 people and growing to 40, dedicated “from morning till night” to platform technology, security, compliance, and tool development. The choice is straightforward: rather than outsource the learning curve, staff the capability so the institution can keep adapting as models, vendors, and use cases shift. Internal capacity also shifts IT's posture from gatekeeper to enabler, enabling the team to ship repeatable components rather than rely on one-off approvals.
That capability created room for a structured way to surface real work. Instead of waiting for a single “killer app,” ASU ran internal grant programs that paired licensing with engineering and developer support. Lev expected “40 or 50 great ideas.” The response was an order of magnitude larger. After multiple rounds, ASU now has more than 600 projects in flight, expanding from faculty-centered experimentation to staff-, student-, and persona-based work, including healthcare-focused experiences and interactive avatars using multimodal tools. Importantly, ASU paired building with “storytelling along that journey,” capturing what teams tried, what worked, and what did not, so the organization can learn in public rather than only through internal slide decks.
Under the hood, Gonick describes parallel tracks: partnering with OpenAI, evaluating other model ecosystems, and building internal tooling. The internal effort is branded as the CreateAI platform and CreateAI builder, aiming to provide a low-code, no-code environment that supports broad adoption. The throughline is pragmatic. Even if you pick partners, “it is still really early days,” so you design for change and keep enough internal leverage to avoid becoming a pure procurement function.
When asked what a new university CIO should do in year one, Gonick’s answer is workflow-first. First, “focus in on a transformational teaching and learning experience” by putting tools in the hands of students and faculty. Second, attack the “incredibly… friction-filled journey” of applying, financial aid, and onboarding, and deliver experiences that feel like the consumer tools students already use. Third, invest in internal IT capacity: “Don’t wait for the market… you will be left doing nothing.”
Gonick closes with an operating principle that fits everything ASU is doing: you do not win by debating AI forever. You win by building, learning, and adjusting. Borrowing from Alan Kay, he adds his own version for this moment: “just start and be ready to pivot.”
