Students create a working AI tool with no coding required
During the course, Losoya built Stitch Finder, a custom AI chatbot GPT that asks users about their sewing experience, their budget and what they hope to make, then recommends a machine and explains why it fits their needs.
The work didn’t begin with code, but with research. Losoya spent weeks combing through sewing blogs and product reviews, compiling machine features, price points and trade-offs into a spreadsheet. Once she had a solid dataset, she uploaded it into CreateAI Builder and began writing prompts that made the bot sound genuinely helpful rather than intimidating.
“I instructed the bot to sound very beginner-friendly and not to use any exaggerated sewing knowledge terms,” Losoya said. That instinct paid off, but it also took trial and error as she kept refining its voice based on feedback from classmates and professors who tested the tool.
What surprised Losoya most wasn’t the sewing research — it was how manageable the technology itself turned out to be. “I thought it was going to be something really extensive and complicated, and it actually wasn’t,” she said.
Losoya wasn’t the only AI in fashion student who created an AI-powered fix for a common consumer frustration. Celeste Polder, a senior at ASU’s Walter Cronkite School of Journalism and Mass Communication, tackled another guessing game: picking the right skincare products. Faced with the flood of products, ingredients and influencer advice that often has little to do with an individual’s actual skin, Polder used CreateAI to develop the Custom Skincare Builder, that delivers evidence-based recommendations instead of guesswork.
Polder pored over peer-reviewed studies on skin aging, acne, oil production and sun protection, along with ingredient research on why certain compounds, such as fat- and water-soluble ingredients, don’t always mix well together. She structured the material so the AI could directly cite sources, replacing vague internet advice with recommendations grounded in science.
She tested the tool repeatedly, catching moments when it echoed the same product suggestions or didn’t ask enough follow-up questions. She then rewrote its instructions until it could guide everyone from total beginners to seasoned consumers.
Where AI is part of the curriculum
Ellis has built her teaching around this kind of experimentation. She launched the 300-level AI in fashion course three years ago through the AI Innovation Challenge and introduced an online 200-level AI foundations for fashion and design course this summer. In this course, lecture videos are followed by conversations with a bot, AI avatars introduce each module, and there is even an AI-generated image detective game. By the end of the course, every student constructs a personal portfolio website showcasing their work.
Ellis has redefined how she measures learning, requiring students to submit transcripts of their AI interactions alongside their final products. “It’s not just checkbox. It’s very much about how the student went through the process.”
Unleashing the next wave of responsible innovation
To date, over 50,000 users have engaged with CreateAI Builder.
With the university’s flagship AI toolkit CreateAI Builder now placing custom intelligence directly in the hands of faculty, staff and students, the next wave of responsible innovation is open to anyone on campus.