How AI Is Transforming Business Operations in K-12

Recently, Matt Penner, Val Verde’s director of information and instructional technology, used Copilot as a code assistant. The AI tool helped him fine-tune a custom application for state reporting that integrates data from different sources, including the student information system. Using Copilot, he finished the task in five hours instead of 10.

“I told Copilot, ‘This is what I want to do. What would you suggest?’” Penner recalls. “What it came up with was phenomenal. I could have come up with something similar, but it would have easily taken twice as long.”

RELATED: Experts say technology can support cash-strapped schools.

More K–12 early adopters like Penner are using generative AI to improve worker productivity and communication and to optimize business operations. They are also using the tools to create or summarize content or to analyze data to discover insights and produce reports.

For example, districts can use the technology to analyze demographic data to develop projections for future growth, while transportation departments use it to create more efficient bus routes, says Tom Ryan, co-founder of the K-12 Strategic Technology Advisory Group and a former CoSN board chair.

“A year ago, people were still afraid of it. Now, we’re getting people to use it,” he says.

To adopt generative AI, districts need to provide training, develop policies that include guardrails to protect data privacy and make sure they align AI with district goals, Ryan says.

Val Verde Schools Reap Early AI Benefits

When Val Verde launched its pilot in 2024, it used Microsoft Copilot because it had standardized on Microsoft 365. It also purchased some ChatGPT subscriptions.

First, the district created an AI committee made up of teachers, administrators, and business and school staff. Then it sent them to training sessions and encouraged them to start using the tools.

Penner also took a lead in teaching departments the basics, such as prompt engineering — how to ask the right questions to get precise information from the AI. He also warns users not to share personally identifiable information and that generative AI tools can give wrong answers, so they need to confirm everything.

DISCOVER: Avoid these 4 common AI mistakes.

Today, about 150 employees are testing the AI tools. While teachers are using generative AI to create more engaging classroom activities, some principals use it to formulate announcements and brainstorm activities to encourage parent night attendance. The district’s HR staff use it to improve job interview questions. Other staff use it to draft board documents, project plans or to improve their email messages, Penner says.

Penner recalls when the risk management director was out of the office all day for district business. When he came back to the office, he asked Copilot to find and summarize the most urgent of his 400 unopened emails, and Copilot whittled it down to a more manageable 37, he says.

“We’re still very much in the early state of doing day-to-day tasks,” Penner says. However, “as the district migrates data center operations to Microsoft Azure, the IT team will increasingly take advantage of Copilot to analyze the cloud environment and ask for network security recommendations.”

Texas School Leader Uses AI to Improve IT Operations

In Texas, the Mount Pleasant Independent School District has also embarked on the initial phases of AI adoption. On the instructional side, teachers have begun using generative AI to develop more creative lesson plans.

While district administrators and staff have clamored to use Copilot, district CTO Noe Arzate has not yet rolled it out for wider use. He plans to deploy it in fall 2025 or early 2026. In the meantime, he’s using generative AI tools to speed software development and make IT operations more efficient.

He has tested large language models and adopted Anthropic’s Claude AI as his go-to tool for technical questions. He recently worked on a project to share data from the student information system with different applications. He specifically wanted to connect a data pipeline using Apache Airflow, but he was having trouble creating a directed acyclic graph — a conceptual representation of activities — to enable the orchestration.

“I gave Claude some context and some parameters, and within a few minutes, I got exactly what I needed to get past the problem,” Arzate recalls. “It saved me probably hours of trying to find the answer online.”

Arzate writes his own software code, but he also uses the Cursor AI code editor to refine his code. Cursor AI has a built-in LLM, but it allows him to switch to Google Gemini and Copilot to see if he gets better results. He recently wrote some code that ran sluggishly, so he turned to Gemini, which revised his code to run faster.

DIG DEEPER: Use these 3 tips to master operationalizing AI.

AI-Powered Chatbots Reduce Help Desk Calls

In Florida, the IT team at Brevard Public Schools used a Microsoft tool to build a generative AI-powered chatbot, and it was so successful at saving staff time that the district plans to build more, says IT Director Barrett Puschus.

The district recently upgraded to a new student information system, and only a handful of staff members knew the technology inside and out. Even though the IT staff posted helpful documents online, they were still fielding a lot of calls for help.

The Brevard team used Microsoft Copilot Studio, an easy tool for building AI chatbots. Then they fed it the help documents to create the chatbot. Now the chatbot answers questions from people throughout the district, which has dramatically reduced calls.

“We’re still early in it, but it’s been effective,” Puschus says. “It’s given time back to staff.”  

The district also plans to build a chatbot for IT help desk support. The tool is cost-effective.

For its generative AI pilot, Brevard Public Schools initially bought 100 Copilot licenses last year and purchased an additional 100 in February. While 80% of users are educators, 20% are business users, including procurement, HR, accounting and IT, Puschus says.

For example, the accounting staff uses Copilot to design complex formulas for their Excel spreadsheets. The procurement department uses generative AI prompts to update its templates for contracts, bids and requests for proposals. Puschus has also begun using Microsoft Security Copilot to more efficiently manage security alerts and automate tedious security tasks.

Overall, generative AI has freed up staff time and improved operational efficiencies, he says. “It’s given joy back to our staff. AI takes away administrative hassle and lets everyone do the things they want to do.”