GigaChat is a generative AI model by Sber designed for learning and creativity, with capabilities to generate videos, text, greeting cards, and music.

We spoke with users to better understand what tasks they accomplish using AI, and if they don’t use it, why.

We explored the competitive landscape across AI-driven products and time-spent platforms, distilling key engagement drivers, interaction patterns, and growth mechanics to shape a more compelling and habit-forming product experience.

We designed a clear and intuitive user flow to guide users from discovery to meaningful interaction with the product. It helps structure the experience, reduce friction, and ensure users can easily understand how to use AI features and achieve their goals.

We conducted two iterations of usability testing using the RITE (Rapid Iterative Testing and Evaluation) method. This approach allowed us to identify issues early and make immediate improvements between sessions. After the first round, we refined key flows and interaction patterns, then validated the changes in the second iteration. This helped reduce friction, improve clarity, and ensure a more intuitive user experience.

The solution led to measurable improvements in key product metrics, including DAU/WAU, time spent, and retention.

The personalized content feed increased user engagement and encouraged more frequent returns, while seamless access to creation tools extended session duration. Clearer value and improved usability contributed to stronger retention and more consistent user activity over time.