Case Study
Case Study
This case study shows how I used AI as a thought partner to make ESL lessons more personalized. Working with a lower-intermediate student, I extracted a segment from a Zoom lesson and created a “revised version” of our discussion using voice cloning. The purpose was not to replace the student’s voice, but to let them hear their own ideas expressed more fluently and clearly.
By comparing the Before Revision and After Revision versions, students could imagine how their speech might sound with greater fluency, turning each lesson into both a practice space and a vision of progress. Click the buttons below to listen.
When teaching ESL, I found that standard textbooks were often too generic and disconnected from students’ real-life needs. For example, one student loved paddleboarding, yet the textbook lesson focused on picnicking. This mismatch made practice feel less relevant and harder to apply.
Learners wanted authentic, personalized practice that could be used in daily life. At the same time, the rise of AI tools sparked my curiosity, leading me to explore how AI could make lessons more efficient to prepare and more innovative in providing realistic, personalized models of English.
To make each lesson more engaging, relevant, and personalized, I set a clear goal: design materials that stay as close as possible to each student’s real-life needs.
My initiative was to work with AI as a partner—analyzing each student’s current level and identifying when and where they would next use English. This could mean preparing for a grocery trip, a church visit, or an upcoming journey. In other words, today’s lesson became a rehearsal for tomorrow’s experience.
I believed this approach would:
Boost confidence by helping students feel ready for real-life interactions.
Increase engagement because students knew they would soon apply what they learned.
Create authentic practice opportunities outside class.
Support retention through paced, purposeful learning.
I recorded one-on-one ESL sessions on Zoom and transcribed them using Microsoft Word’s Dictate feature. Then, with NotebookLM, I analyzed transcripts to spot patterns, identify areas for improvement, and extract the exact sections needing attention. This saved time and gave me precise insights into each student’s speaking needs.
I focused on student-led discussions—such as trips or favorite hobbies, where they applied vocabulary most naturally. Using ChatGPT, I transformed excerpts into fluent, native-like versions. This sparked student curiosity by showing them how their own speech could sound more natural and confident.
To bring lessons to life, I used Genny to generate realistic audio of the revised dialogues, cloning both my voice and the student’s (with their consent). This innovation created personalized audio models, helping students hear and imagine themselves speaking fluently in authentic conversations.
I built vocabulary tables directly from transcripts with ChatGPT, including definitions and examples. After each lesson, I packaged the revised transcript, audio dialogue, and vocabulary table into a single handout, an efficient system for home practice and review.
Students gained personalized, realistic models to practice speaking with confidence.
Vocabulary retention improved through targeted review tables.
Learners reported higher engagement and motivation, knowing the lessons tied directly to their real-world contexts.
This case study demonstrates my ability to:
Conduct a needs analysis using authentic learner data.
Apply AI tools for efficiency in content creation and personalization.
Spark learner curiosity by designing lessons around their real-life contexts.
Leverage innovation through multimedia integration (AI voices, transcripts, custom handouts).
Together, these showcase my instructional design mindset even while working in a volunteer ESL role.