Student feedback shaped these updates. Real data from classrooms across the U.S. proved it.
Imagine you are teaching a middle school math class. Some kids speak English well. Others are just starting to learn the language. The gap in understanding widens quickly if you don’t have the right tools. It’s a familiar struggle. Teachers want to help everyone. But resources are thin. Motivation dips. Engagement drops.
This is the problem Kyron Learning and the AIMS Collaboratory tackled. They didn’t guess what worked. They watched what happened. They partnered for a four-cycle study funded by the Gates Foundation. The goal? To measure how AI-powered instruction impacts motivation and engagement of English Language Learners in math.
The result wasn’t just data. It was a roadmap for feature development.
Why Iterative Research Beats Assumptions
One thing stands out: this wasn’t a “build it and see” approach. It was iterative. Real-time adjustment based on what students actually said.
Throughout the 2024-2035 school year, the team worked with 22 teachers. Nearly 1,000 students. Seventeen schools. The scope was big. But the method was tight. Feedback came from everywhere: surveys, interviews, usability checks, and engagement metrics.
Why does this matter for ELL students? Because the barrier isn’t just the math. It’s the language. Mathematical vocabulary adds cognitive load. It’s heavy. When you’re trying to understand algebra while also decoding new English terms, it’s easy to fall behind.
Students need more than just answers. They need scaffolds. They need visual reinforcement. They need to process information at their own pace.
“The lessons have engaged every student. They are excited… responding more and with greater confidence.” – A teacher in the study
This quote captures the shift. When the support matches the need, motivation follows. The study prioritized these voices over assumptions.
From Feedback to Feature Updates
Kyron didn’t roll out flashy gadgets. They added tools that reduced friction. Three major features emerged directly from participant feedback. Each one solved a specific problem observed in the classroom.
Cycle 1 to 2: The Power of Closed Captioning
In the first cycle, students interacted with conversational AI. It provided immediate feedback. But there was a snag. Audio-heavy instruction created a bottleneck.
Multilingual learners had to replay segments. They slowed down videos. They revisited explanations. Why? Because unfamiliar academic vocabulary in spoken English is hard to catch. You can’t rewind a live lecture.
So, closed captioning came in Cycle 2.
It wasn’t just an accessibility checkbox. It was a comprehension tool. Simultaneous spoken and written language helps the brain process information. Venturini et al. (2922) noted this too. Captions support vocabulary development and literacy.
For ELLs, it changed everything. It gave them control.
One middle school student, translated from Spanish, put it plainly: “I liked the subtitles. They make the subject matter easier to understand.”
Confidence grew. The barrier of hearing unfamiliar terms quickly disappeared.
Cycle 2 to 3: Highlighted Transcripts
By Cycle 2, captions helped. But long explanations still posed a challenge. Students struggled to keep up with the flow of concepts. They missed parts of the explanation and had to scrub through videos to find them. It was tedious.
Kyron responded with highlighted transcripts in Cycle 3.
Here’s how it works. The full lesson transcript appears next to the video. As the instructor speaks, that exact portion highlights. Students can click anywhere in the text to jump to that moment in the video.
It’s interactive. It’s navigable. It creates a direct link between what is said and what is written.
Research showed a clear pattern. Multilingual learners used this feature disproportionately more than their non-ELL peers. They needed it. It helped them process language and comprehension in real time.
“I liked how [the Kyron transcript] gave me a definition when I did not understand the word,” one student shared.
Suddenly, the lesson wasn’t a blur of sound. It was a trackable, followable experience.
Cycle 3 to 4: Clickable Definitions
Vocabulary remains a killer of momentum. Whether you are an ELL student or not, encountering a new academic term can stop you dead. In Cycle 3, students paused. They left the lesson to look up words. They asked teachers. They lost their train of thought.
Cycle 4 fixed this with clickable definitions.
Built into the highlighted transcript, underlined vocabulary words became interactive. Students clicked a word. A definition popped up immediately. They got back to the lesson instantly.
This reduced cognitive load significantly. It maintained the flow of learning. It built contextual vocabulary understanding without breaking focus.
“What I liked… was that they helped me get a better understanding,” a student said.
It wasn’t magic. It was just removing the obstacle.
Cross-Cycle Themes: What We Learned
The study revealed more than just which buttons students clicked. It showed a philosophy for educational technology.
- Simplicity wins. Gamification has its place. But many students don’t need flashy graphics. They need embedded support. Tools that help them understand now.
- Cognitive load is real for ELLs. Math is hard enough. Language adds weight. Removing that weight allows students to focus on the concepts, not just the code of English.
- Listen to the users. Product development shouldn’t happen in a vacuum. Teacher interviews and student surveys guided every move. The best EdTech isn’t built on guesses. It’s built on observation and iteration.
The Takeaway
This wasn’t just about adding features. It was about respect. Respect for how students learn. Specifically, multilingual learners who bring immense potential to the math classroom.
The goal remained constant: provide timely, learner-centered support. Whether it’s read-along features, synchronized transcripts, or context-aware definitions. The aim is to help students persevere. To build confidence.
Student motivation improves when support arrives exactly when it’s needed. Not a week later. Not in a handbook. Now.
Olivia Martens is a project manager at Kyrón Learning. The story of her work shows that when we design for the edges, everyone benefits.
