Schools and districts are investing time and resources in artificial intelligence training, but an important question remains: What happens after the workshop ends?

Teachers may leave a professional development session knowing how to write a prompt or use a particular AI tool. But knowing how a tool works isn’t the same as knowing when to use it, how to incorporate it into instruction or how to help students learn with it responsibly.

West Virginia is developing a statewide approach designed to bridge that gap.

In a recent episode of the Making Schools Work podcast, Cliff Sullivan, AI coordinator for the West Virginia Department of Education, joined SREB’s Daniel Rock and Leslie Eaves to discuss how the state is building an AI professional learning system that moves educators from foundational knowledge to classroom application and collaborative problem-solving.

The approach offers lessons for school and district leaders looking to make AI professional learning more meaningful, practical and sustainable.

Start With Teachers’ Needs, Not the Technology

Before joining the West Virginia Department of Education, Sullivan worked with the West Virginia STEAM Technical Assistance Center, helping educators integrate hands-on learning into their classrooms.

Rather than simply demonstrating a lesson and leaving, his team worked alongside teachers to develop activities connected to their existing standards and instructional goals.

That experience shaped his approach to AI professional learning.

Sullivan found that teachers wanted more than information about new technologies. They wanted someone to work with them, help them understand how to apply new ideas and develop resources they could actually use.

West Virginia’s current approach builds on that lesson. Instead of asking teachers to adopt a predetermined set of AI activities, the state is designing professional learning around educators’ needs and the challenges they encounter in their schools.

The principle is straightforward: Professional learning should produce something teachers can use, not just something they can remember.

For school leaders, that means starting with questions such as:

• What instructional or administrative challenges are our educators trying to solve?
• Where could AI support work teachers are already doing?
• What do educators need to know before using AI responsibly?
• How will teachers have opportunities to practice, collaborate and refine what they create?

These questions can help shift professional learning from introducing technology to improving practice.

Build a System That Can Reach Every School

One of West Virginia’s biggest challenges is scale.

With 55 counties and thousands of educators, Sullivan recognized that traveling from school to school to deliver individual training sessions wouldn’t be enough.

The state’s developing solution is a train-the-trainer model.

Regional specialists will support local leaders known as AI Mountaineers, who will help deliver professional learning within their counties and schools. Educators will also have access to asynchronous learning opportunities to deepen their understanding of AI concepts and tools.

The approach is designed to build local expertise rather than rely on one person or a small state-level team to provide all the training.

It also recognizes that different districts use different technologies. Although West Virginia uses Microsoft at the state level, some counties have adopted other platforms. The professional learning model emphasizes transferable AI knowledge and skills rather than requiring every educator to use the same product.

For district leaders, the broader lesson is that sustainable professional learning requires both a shared direction and local flexibility.

A district might establish common expectations for AI literacy, data privacy and responsible use while allowing schools and teacher teams to explore applications that meet their particular needs.

Give Educators a Path From Awareness to Application

West Virginia’s professional learning initiative is called PAINT, which stands for Pedagogical AI for New Teaching.

The name reflects the initiative’s focus on helping educators rethink what teaching and learning can look like with AI.

PAINT includes hands-on professional learning experiences in which teachers explore tools, collaborate with colleagues and develop resources connected to their own classrooms.

The broader statewide plan uses a mountain-themed progression with three levels of learning.

Pioneer: Build foundational AI literacy

Educators begin by learning basic AI concepts, exploring generative AI and developing prompting skills. They also consider responsible use, ethics and practical classroom applications.

The goal is to help educators become comfortable enough with AI to make informed decisions about how it might support their work.

Trailblazer: Apply AI to real challenges

At the next level, educators move beyond experimenting with individual tools. They collaborate to develop projects, create resources and explore AI-supported solutions to classroom or grade-level challenges.

This might involve designing a more engaging learning experience, creating a tool to support a recurring task or developing a cross-curricular project.

Ridge Climber: Solve larger problems together

At the highest level, educators work collaboratively on challenges affecting an entire school, department or educational system.

Sullivan describes possibilities such as addressing chronic absenteeism, improving scheduling or helping students explore career pathways.

The progression is designed to help educators build their skills over time rather than expecting them to master everything in a single training session.

It also aligns closely with SREB’s AI Professional Learning Framework, which identifies four phases of professional learning: awareness, exploration, application and iteration.

As Eaves explains in the podcast, AI professional learning is not something schools can complete in a day. It requires educators to develop their understanding, practice using AI, reflect on the results and continue refining their approaches.

Make Professional Learning an Opportunity to Practice

Sullivan’s approach to professional learning includes a simple but effective strategy: Put teachers in small groups around one device and give them a problem to solve.

One educator operates the device while the others contribute ideas, discuss prompts and evaluate the results.

This creates opportunities for teachers with different experience levels to learn from one another.

A teacher who has never used AI can participate alongside someone who is more experienced. Together, they can test ideas, examine what works and discuss how a particular application might fit their classrooms.

The activity also makes the learning collaborative rather than passive.

For example, a professional learning session could ask teachers to select an existing lesson and explore how AI might help them differentiate instruction, strengthen questioning or create a more engaging activity.

Rather than leaving with a list of suggested tools, teachers leave with a resource they helped create and a better understanding of the decisions behind it.

For school leaders planning professional learning, this offers an important reminder: Give educators time to do the work during the training, not just hear about how they might do it later.

Focus on Screen Value, Not Just Screen Time

A particularly important part of West Virginia’s approach is its emphasis on what Sullivan calls “screen value.”

The goal isn’t necessarily to put more technology in front of students.

Instead, educators are encouraged to consider whether AI can make learning more meaningful, even when students aren’t using a device themselves.

For example, a teacher might use AI during planning to develop a hands-on project, create discussion questions or adapt an existing lesson to meet different student needs.

Students could then complete the activity through collaboration, experimentation, discussion or other experiences that don’t require additional screen time.

In other cases, students might interact directly with an AI tool because it creates an opportunity for deeper learning.

The distinction is whether technology is contributing something valuable to the experience.

Before introducing AI into a lesson, teachers can ask:

• Does this activity require students to think, create or solve problems?
• Is AI helping students engage more deeply with the content?
• Would the learning experience be stronger with or without the technology?
• What will students understand or be able to do afterward?

These questions keep instructional goals at the center of decisions about AI use.

Use AI to Challenge Students, Not Give Them the Answers

One of Sullivan’s most interesting classroom examples involves a history lesson.

Imagine students studying the American Revolution. Instead of asking an AI tool to summarize the war, the teacher creates an AI agent that argues the British won.

Students know that conclusion is incorrect, but they must use primary-source evidence to challenge the AI’s claims.

The agent is designed to continue questioning students rather than simply providing the correct answer.

Students must examine documents, develop arguments, cite evidence and defend their conclusions.

In this scenario, AI doesn’t remove the thinking from the assignment. It creates a challenge that requires students to think more carefully.

The teacher can then build on the activity through classroom discussion, giving students opportunities to explain their reasoning and reflect on the evidence they used.

It’s an example of the difference between using AI to complete an assignment and using AI to support learning.

The same principle can apply across subjects. Teachers might use AI to generate a flawed mathematical argument for students to analyze, present competing interpretations of a literary passage or challenge students to evaluate proposed solutions to a scientific problem.

The important element isn’t the technology itself. It’s the thinking students must do in response.

Start Small and Build Capacity Over Time

West Virginia’s statewide initiative is ambitious, but its approach doesn’t require every educator to become an AI expert immediately.

Sullivan emphasizes starting with manageable goals, developing useful applications and building toward more complex work.

That principle applies just as well to an individual school or district.

A school might begin by helping teachers understand approved AI tools and responsible-use expectations. Teacher teams could then identify one recurring instructional challenge and explore whether AI could help address it.

As educators gain confidence, they can share what they’ve learned, refine their approaches and collaborate on larger challenges.

Over time, those small applications can contribute to a broader culture of thoughtful AI use.

West Virginia’s model is still being implemented, and its long-term outcomes remain to be seen. But its design offers a valuable example of how schools can move beyond isolated training sessions toward a sustained approach to professional learning.

The goal isn’t simply to teach educators how to use AI. It’s to help them develop the skills, judgment and confidence to decide when AI can make teaching and learning better.

Hear the Full Conversation

In the full Making Schools Work podcast episode “Beyond One-and-Done: How West Virginia Is Scaling AI Professional Learning,” Sullivan, Rock and Eaves discuss West Virginia’s PAINT initiative, its statewide professional learning structure and practical ways educators can use AI to strengthen instruction in more depth.

Make sure you catch every episode by liking and subscribing to the Making Schools Work Podcast wherever you listen to podcasts.