How 3 California School Business Leaders Are Putting AI to Work

Artificial intelligence (AI) is rapidly moving from theory to practice in California school districts as leaders increasingly embrace it — not as a futuristic concept, but as a practical tool for improving operations, communication and efficiency. At CASBO Con 2026, several school business leaders shared how they are exploring AI to support everything from board communications to data analysis to workflow automation. Grant Schimelpfening, assistant superintendent of administrative services at Lindsay Unified School District, has integrated AI into operational and strategic work, particularly around data analysis, communications and planning.

“From the operations perspective, the main challenge was making better use of large amounts of fiscal and operational data,” he says. “Turning that information into something actionable can take a significant amount of staff time, which limits the time available for strategic thinking and problemsolving. AI gave us a way to move faster from raw data to insight.”

In El Dorado Union High School District, assistant superintendent of business services Bob Whittenberg has emphasized thoughtful exploration, governance and responsible implementation through a districtwide AI Guiding Coalition. “At this stage, our initiative is less about one specific platform and more about building a responsible AI culture,” he says. “We are trying to understand how AI can help people do their work better, while also creating shared expectations around accuracy, privacy, bias, academic integrity and appropriate use.”

At Santa Clara Unified School District, Director of Technology Joe Ayala has focused on building an AI learning culture across departments, helping staff develop practical skills and even create custom workflow tools. “As a tech director, you develop a built-in instinct to automate or script your way out of repetitive work,” he says. “Free up that time and you free up leaders to actually lead.”

While their approaches differ, these forward-thinking leaders share a common philosophy: AI works best not as a replacement for professional expertise, but as a tool that frees educators and administrators to spend more time on the human side of leadership.

Below, they respond to our questions about how they’re putting AI to work in their districts.

In your district, how have you built buy-in among those who are hesitant about using AI?

Whittenberg: We started with simple, practical demonstrations. Once people see AI solve a familiar problem, they become more open to exploring deeper uses. My own growth with AI came from watching more advanced users share how they were using it. Buy-in grows when staff can learn from colleagues, see real examples and experiment in a low-pressure environment.

Ayala: First, we built hands-on workshops using real district workflows on real district documents. People don’t believe a slide deck about AI. They believe what they just did themselves in 20 minutes. Second, we did our best to take the fear out of using it. Hesitation rarely comes from a lack of information. It comes from a fear of looking foolish in a new medium. Address the fear and the rest moves faster. Finally, we put the skeptics on our AI Advisory Committee. We asked them what they were worried about and built our policy around their concerns first. The skeptics became the validators.

Schimelpfening: I talk about AI openly and use it transparently. When I present to staff or other groups, I often show how AI helped me do my job better. Just as important, I have tried to frame AI as a support tool rather than a replacement for expertise or judgment. That has been one of the most important messages: AI is most valuable when it helps create more space for the human side of leadership, not less.

What steps have you taken to ensure sensitive data remains secure and compliant while using AI tools?

Ayala: Our security model evolved in three steps. Step one was anonymization training on free tools. We started by teaching people to never put real names or identifying data into a free tool. That training gave staff something they could do safely on day one without waiting for a procurement cycle. Once we had real demand, we adopted licensed versions of Gemini, ChatGPT and Claude – the editions with contractual data protection, no-training-on-our-data clauses and enterprise data handling.

That removed the largest risk vector from daily use. Every AI tool now under consideration for district use goes through a privacy and data-handling review that addresses data residency, FERPA and COPPA alignment, training data policies and incident response posture. If a vendor can’t answer those questions clearly, they don’t move forward. The pattern matters more than any single policy: Meet people where they are, give them a safe move now, and build the formal infrastructure underneath as adoption catches up.

Schimelpfening: Our district has adopted AI policies that outline expectations for appropriate use, and our IT department has implemented safeguards that allow us to maintain strong oversight. We have also established an AI task force with a variety of stakeholders to help with the ongoing monitoring, guidance and education related to AI use and district policies. That has been especially important because the AI landscape is evolving so quickly.

How are you thinking about equity when implementing AI?

Whittenberg: We are thinking about equity as a central design principle, not an afterthought. For staff, that means making sure AI is not limited to early adopters or those who are already comfortable with technology. Otherwise, AI could unintentionally widen gaps between those who know how to use it well and those who do not. Ultimately, our approach is that AI should be used to level the playing field, not create a new advantage for some and a new barrier for others. That means thoughtful access, clear policy, staff training, student instruction and ongoing monitoring to make sure AI supports our broader goals of fairness, opportunity and student success.

What tangible results have you seen from AI use in your district so far?

Schimelpfening: The most tangible result has been time savings. AI has helped reduce the amount of time needed to review large data sets, work through complex contracts, refine communications and develop plans or presentations. In many cases, it helps move work forward much faster by creating a strong starting point that can then be reviewed and improved. Just as importantly, the time savings have created more space for me to be present in the work itself. That additional perspective has been valuable in helping me make better decisions and remain grounded in the day-to-day realities of the district.

Ayala: Leveling up. The most consistent thing I hear from staff is some version of, “I don’t know if I could have done this without AI.” About a year ago, I learned to code an iOS app in eight hours. Two weeks ago, I built a Google Apps Script during the closed session portion of a board meeting. People are combining new capability with existing expertise to do things that would have taken weeks before, or that they wouldn’t have attempted at all. That’s the real result – the expansion of what one person can accomplish.

What is one mistake or unexpected challenge you encountered — and what did it teach you?

Schimelpfening: The most glaring mistake I made with AI, and something I still have to check myself on, was trusting it too much. In the beginning, I leaned so heavily into the time savings that I made false assumptions about the accuracy of what it was producing. That experience taught me a lesson I have not forgotten: AI can sound confident, polished and convincing, while still being wrong. We still need to verify sources, confirm accuracy and make sure there is an appropriate human review process in place.

Whittenberg: One unexpected challenge was realizing how quickly people move to the extremes when talking about AI. Some see it as a threat that will lead to cheating, job replacement or loss of critical thinking. Others see it as a magic solution that can immediately transform instruction and operations. Both reactions miss the real work. The real challenge is building the culture, guidance and professional learning needed to use it responsibly. Implementation cannot just be a technology decision. It has to involve instruction, operations, ethics, data privacy, equity and community trust.

What advice would you give to another district looking to start or scale an AI initiative with limited resources?

Ayala: Don’t start by buying – use what you already have. Google Workspace and Microsoft 365 both have AI built in. Most districts can gain their first wins without a single new contract. Pick one group, one workflow, one win. Ask the group what they hate doing, build the end-to-end solution, document it, then move to the next group. Trying to roll AI out to everyone at once is how districts end up with shelfware. When you do buy, push hard on the vendors. They want user counts right now and will discount aggressively if you ask. Some AI vendors offer plans where you pay for actual consumption rather than per seat. For districts with uneven adoption curves, that can be far cheaper than buying a license for every employee.

Schimelpfening: Find a few AI champions in your district who are genuinely interested in it and give them room to explore, experiment and produce real work with it. Then have them share – not just what they created, but how using AI has made their jobs better and what they are now able to do with that additional time to better serve students, staff and the community. One of the great things about the school business world is that people are willing to share. There are districts that already have thoughtful systems, guidelines and policies in place. Ask questions, learn from what others have built and adapt those ideas to fit your own context.

Whittenberg: My advice would be to start small and stay mission-driven. It can begin with a small group of thoughtful people asking the right questions. What problems are we trying to solve? Where could AI responsibly save time? Where could it improve access, communication or learning? And what guardrails do we need before we move too quickly? With limited resources, I think the best approach is to identify a few high-value, low-risk use cases. Those early wins can build confidence and show staff that AI is not about replacing people; it is about helping people work more effectively.

In your experience, where does AI add the most value — and where is human judgment still essential in school business operations?

Ayala: I work from a 10/80/10 framework. The first 10% – defining the work, framing the problem, knowing what “good” looks like – is human. The middle 80% – drafting, summarizing, restructuring, surfacing patterns, generating options – is where AI adds enormous value. The final 10% – judgment, review, contextual fit, relationships, political read – is human again. AI adds the most value in prep work, document synthesis, draft generation and pattern recognition across volumes of information no human has time to read carefully. Human judgment remains essential anywhere relationships, politics, personnel decisions, equity calls, board navigation or community trust are in play. The cost of getting one of those wrong is not measured in time saved.

Whittenberg: AI adds the most value in areas where we are trying to organize information, improve efficiency, draft communications, analyze options or move from a blank page to a solid starting point. But AI does not understand our local context, our community history, our board priorities, our bargaining environment or the real-world consequences of a decision unless we provide that context and review the output carefully. It can summarize information, but it cannot determine whether the information is accurate, legally appropriate, fiscally responsible or politically sensitive. In school business operations, the stakes are too high to treat AI as the final authority.

What opportunities or risks are you watching most closely?

Whittenberg: The risks we are watching most closely are data privacy, accuracy, bias, academic integrity, overreliance and equity of access. We do not want AI to widen existing gaps or replace the professional judgment of educators and staff. Ayala: Data exposure is always a concern. External pressure is another – from hackers, phishers and people overloading our systems with AI-generated lawsuits and California Public Records Act (CPRA) requests no records team is staffed to handle. One area I’m focused on is how we counter AI bots and agents with our own set of AI agents. The defensive perimeter is going to need its own AI.

If you could implement one AI capability tomorrow with no barriers — budget, policy or technical — what would it be and why?

Ayala: Just one? The first would be an AI-native Student Information System (SIS) and financial system that lets you build reports and pull data using plain English. How many English-learning students do I have? Are they on track to graduate? What’s the balance of my budget? And eventually, the system surfaces alerts on its own when a student is at risk of failing a class or an account is trending over budget. The second would solve the “I didn’t know that” problem. A modern district intranet connected to a districtwide retrieval system where any staff member could ask a plain-language question and get the right answer pulled from board policy, ed code, the union contract, prior board actions, vendor agreements and historical district communications – with citations to the source. Knowledge management is the silent crisis in K-12. Every district has the answer somewhere, yet almost nobody can find it. Solve retrieval and you unlock more time, better decisions and better continuity through staff transitions than any other single intervention I can think of.