How one teacher used AI to streamline student feedback

Julia Tingey, a forensic science instructor at Canyons Technical Education Center, kept running into the same challenge in her classroom.

Her students were submitting electronic discussion posts for forensic investigation assignments, but meaningful feedback often took days to reach them. Between reviewing submissions, grading, and responding to students in Canvas, the turnaround time could stretch to 3–5 days.

“I spend two to three weeks grading student work or my students’ discussion posts, and I also have to interact with the students about their posts in Canvas, so that takes up another week,” Julia explains.

By the time students received feedback, the class had often already moved on to something new. Students who needed more support were often left waiting the longest, and opportunities to improve their work and understanding in the moment were easy to miss.

So Julia started exploring another approach.

Working alongside researchers from Utah State University’s Center for the School of the Future, she used MagicSchool to build a custom AI-powered learning scaffold designed to give students immediate, guided feedback while students were still actively working on their responses.

The Center for the School of the Future later documented the experience in a research practice brief, examining how the educational chatbot shaped student learning, classroom engagement, and the feedback process itself.

A classroom tool built with clear guardrails

Using MagicSchool, Julia and researchers from Utah State University Center for the School of the Future created a custom AI chatbot for the students of her 11th-grade forensic investigation course. Students used the tool while drafting electronic discussion posts tied to a “Forensic Investigation” assignment. From the beginning, the experience was intentionally structured around instructional guardrails and student ownership. The chatbot relied on Retrieval-Augmented Generation (RAG), meaning responses were grounded only in approved course materials rather than the open internet.

The AI chatbot guided students to answer every part of a question and expand on their reasoning, while still leaving the final writing to the student. It couldn’t generate final responses for students, but summarized students’ thinking back to them so they could continue refining their own work. Julia was also transparent with students about how the tool worked and why they were using it in class. As she shared in the practice brief, “I developed this AI chatbot to keep learning transparent, rather than secretive.”

What changed when student feedback came faster?

Fifteen students in Julia’s 11th-grade forensic investigation course used the educational chatbot while developing electronic discussion posts for class assignments. Instead of waiting several days for teacher feedback after submitting their work, students received feedback and guidance while their ideas were still taking shape. According to the practice brief, many students became more thorough in their initial responses because they knew they would receive immediate feedback and follow-up questions as they worked. 

One student shared that the AI chatbot helped them “think about the question from a different angle,” while another appreciated being prompted “to go into more depth” in certain parts of their response.

Students also used the chatbot differently depending on their needs. Strong writers often treated it as a quick check, while students who needed more help leaned on it more heavily to organize and strengthen their ideas. Because AI use was transparent and expected within the assignment, the experience also reduced anxiety and ambiguity around when and how students could use AI responsibly in the classroom.

What did researchers learn?

Researchers at Utah State University Center for the School of the Future found that with Julia’s approach to the educational chatbot, students generally produced stronger initial drafts and more complete final electronic discussion posts as part of the assignment process. The most immediate shift came from timing. A student feedback cycle that previously stretched across several days became far more immediate, while students were still actively working through their ideas.

The research also reinforced the importance of strategically pairing teacher leadership and oversight with the AI chatbot. Julia remained fully responsible for instruction, assessment, and grading, while the educational chatbot supported students with feedback during the drafting and reflection process. The practice brief also noted that students engaged with the tool differently based on their confidence and needs, allowing the scaffold to feel flexible rather than one-size-fits-all. 

Researchers documented areas for improvement, including moments when the chatbot oversimplified student responses or felt unnecessary to some learners. Including those challenges helped paint a fuller picture of what responsible classroom implementation actually looks like in practice.

From one classroom to broader adoption

What Julia and the researchers at Utah State University Center for the School of the Future explored in a single forensic science classroom points to something larger for schools and districts thinking about AI adoption. The model itself was relatively simple: one teacher, one platform, one assignment rubric, and a clearly defined set of instructional guardrails. Yet the introduction of the AI chatbot created space for faster feedback, stronger student reflection, and more support during the learning process.

The practice brief notes that the same foundational elements—transparency, guided support, grounded course materials, and teacher oversight—can travel across subjects, grade levels, and schools. It also reinforces how thoughtful AI implementation starts with teaching and learning goals, not with the technology itself. 

For schools exploring how to bring AI into teaching and learning responsibly, Julia’s classroom offers a practical example of what that can look like. The technology was only one part of the story. The instructional design behind it made the difference, with clear guardrails, transparency with students, grounded course materials, and a teacher who remained fully in control of learning, feedback, and assessment throughout the process.

The work documented by Utah State University Center for the School of the Future also highlights how timely feedback can change how students engage with their learning. When students can receive feedback while they’re still engaged in the assignment, they have more opportunities to revise and strengthen their work before the learning moment passes.

Interested in learning more about Julia’s work at Canyons Technical Education Center? Read the full practice brief from Utah State University’s Center for the School of the Future.