Junier is an AI tool I designed, built, and launched independently - from initial concept through to final product. It helps teams build better products with feedback-powered insights for whatever they are working on.
I started Junier because I saw a fundamental change in how teams do product work due to AI, and I believed there was an opportunity to build a tool for this future.
See more at https://junier.co.
Building Product Knowledge
A core technical challenge was generating unique, product-specific data powering contextual insights for teams. This starts with Sessions, where users run sessions with participants using their own prototypes or live product. Junier creates and stores knowledge about the product by connecting what happens on screen to what participants say and do. That data then drives real-time insights wherever the designer is working, using image and text data to infer work context and surface relevant feedback.
This involved a complex AI pipeline: recording and storing session data (audio, video, screen), multi-modal AI classification, LLMs, embeddings, a vector database, and RAG.
