- Learning & Research
- ~ 80%
- Work & Productivity
- ~ 60%
- Coding
- ~ 40%
KIMI K3 Mobile Experience Upgrade
Making complex tasks easier to start and continue
K3 upgraded Kimi’s mobile experience for complex AI tasks. I supported the mentor-led redesign and independently analyzed 502 survey responses, producing a research dashboard and report.
01 / THE PROBLEM
Understanding friction in complex AI tasks
Analysis for competitive products
Product Name
Top 3 usage
- Depth Reasoning
- Coding
- Multimodal
- Coding
- Depth Reasoning
- Tool Ecosystem
- Depth Reasoning
- Coding
- Multimodal
- Depth Reasoning
- Multimodal
- Cross-Product Comparison
- Depth Reasoning
- Coding
- Multimodal
Insights
Users Already Rely on AI for Complex Tasks
Mobile Needs to Support the Complex Tasks Users Already Perform
The Experience Lacks Clarity and Predictability
Users need clearer task progress and expectations, while interface improvements should not be framed as solutions to model quality or speed.
- Inconsistent Results & Poor Intent Understanding
- ~50%
- Long Wait Times
- ~ 40%
- Usage Limits
- ~ 40%
The organization of capabilities needs improvement.
Users don’t understand the purpose of certain features, can’t find the features they need, and struggle to understand the relationship between Agent entry points and how to access history.
Keep simple interface
Many users prefer a simple and clear interface, suggesting that more capabilities need to be organized while maintaining clarity.
02 / KEY EXPERIENCE
K3 Mobile Experience
01 · Start the right task
Clarify the entrance and enhance the user input experience.
02 · Know what is happening
Clearly explain each model and provide recommendations. Notify users when switching models and when they are approaching their usage limits.
03 · Pick up where you left off
Improve conversation history management and search, while preserving saved prompts for quick access and efficient workflows.
03 / REFLECTION & CONTRIBUTION
Reflection & Contribution
Reflection
Although the redesign introduced more status and progress cues, it did not address the underlying issue of long wait times. Users continued to report excessive delays, which may indicate an imbalance in how tasks are allocated across different reasoning levels.
In addition, token usage and consumption remain difficult for users to perceive. Without clear quantitative indicators, users lack a concrete understanding of how much they have used, how much remains, and how different tasks affect their usage.
Contribution
I independently analyzed 502 survey responses and created the dashboard and research report. I contributed to the mentor-led K3 mobile redesign as a supporting designer.