In the initial phase of user research, I worked closely with the product team to conduct user interviews with several translation teams from our B2B clients across industries such as e-commerce, banking, pharmaceuticals and national security. The goal was to gather insights into their unique needs, workflows, and pain points surrounding speech-to-text translation.

Through these interviews, we explored how translation teams currently handle audio and video files, what challenges they face with existing tools, and what features would make their processes more efficient. We uncovered that teams were often struggling with manual transcription for audio and video content, leading to delays and inconsistent results. Translators also expressed a need for more accurate speech recognition in different accents and languages, as well as seamless integration with  CAT tools. Facilitating the use of our advanced linguistic tools like User Dictionaries and Neural Fuzzy Adaptation within the speech-to-text workflow was also key.

The design process for the speech-to-text translation feature was iterative, progressing from low to high-fidelity designs while incorporating continuous feedback from both users and the development team.

I began by sketching out initial concepts and low-fidelity wireframes, focusing on the core functionality and user flows, such as how users would upload audio/video files, initiate the translation process, and manage their media. These early designs prioritised simplicity and clarity, ensuring that the interface was intuitive and easy to navigate.

Next, I moved to high-fidelity mockups, where I refined the visual design, incorporated branding elements, and worked to ensure a cohesive user experience across various screens. During this stage, myself and two product managers also began collaborating closely with the developers through tech framing sessions to discuss the feasibility of specific features. Their feedback on performance, integration, and technical constraints helped refine the designs, ensuring they were not only user-friendly but also technically viable.

Due to resource and time limitations, the new speech-to-text feature would share the same base component as our pre-existing text file translation feature. This led to tight constraints specifically when designing the post-edition mechanism since the layout for post-editing video files had to keep the same page layout as text files. However, it meant we were able to bring much needed improvements to the post-editor component for all file types.

In the second round of research, we conducted usability testing with a group of target users to gather feedback on the initial designs. This round was crucial for identifying pain points and refining the user experience before moving to development. We observed how users interacted with the feature, paying close attention to their workflows and reactions to the proposed interface.