Problem
A startup developing an innovative chip needed to create firmware and supplementary software to optimize its performance. Beyond firmware, the investment also required identifying a blue ocean opportunity in the software space—specifically, leveraging machine learning to enhance the chip’s market presence.
Competitive Analysis
In the early stages, a detailed competitor analysis was conducted to map out existing market features and user demographics. This helped define the unique value proposition and uncover gaps that could be addressed with novel software solutions.
UX Research
To refine the software’s direction, we conducted interviews with a targeted group of seasoned machine learning experts across various industries. This research helped identify key user needs and workflow challenges. A crucial distinction was made between different types of machine learning applications and their respective workflows.
We deconstructed existing solutions, from platforms like Lobe to open-source machine learning tools, while also engaging in Fast.ai courses to pinpoint differentiating factors for the startup’s unique market entry.
Process & Summary
Through deeper exploration, conversational design emerged as a strategic approach to enhance user workflows. A modular design system and interactive conversational scenarios were developed to guide users through each step of the learning process.
Most machine learning platforms rely heavily on terminal-based workflows. However, by integrating a conversational interface, we aimed to optimize processes and make the software accessible to a broader audience.
The application leverages unsupervised learning to adapt to user needs through conversational interactions. Designing components for this chat-based system posed a challenge due to the hypothetical nature of use cases. To address this, we created an initial persona-driven prototype to better understand software requirements and user expectations.
Interactive prototypes were developed for stakeholders, with iterative user research guiding the lean, agile development process. Highly interactive elements were primarily driven by conversational chat and settings, while machine learning tooling prototypes served as an intermediary step. This helped determine which tools should be seamlessly integrated and which should remain as static external components.
Project
MVP Design & UX Research
Client
Thinci
