Designing an AI App for Effortless Scheduling
Darma is a smart scheduling assistant that automates meeting coordination, resolves conflicts, and helps users manage their calendar using multimodal inputs like voice, text, and snapshots.
My role was to take the core Darma concept and design a native iOS app that feels seamless and intuitive.
Project duration
8 months
Team
Prodcut Designer, Project Manger, Machine Learning Specialist , Front-End Developer
KPI
Volume of events scheduled through Darma
CHALLENGE
Designing a Smarter, Simpler Way to Manage Your Calendar.
Modern professionals spend an overwhelming amount of time managing their calendars - juggling meetings, rescheduling conflicts, and coordinating across time zones. Even with smart calendar tools, the process often remains manual, requiring constant back-and-forth communication. This not only wastes valuable time but also adds mental load to an already busy workday. The challenge was to design an app that could take over the tedious parts of scheduling, intelligently resolving conflicts and organizing the user’s day using the power of AI.
MY ROLE
Designing an intelligent iOS app that balances AI automation with user control and native simplicity.
I worked closely with the product team to shape the product roadmap, define priorities, and design the most impactful features for launch. My focus was on establishing a clear, consistent design system while leveraging native iOS components wherever possible to create a seamless user experience. Because the app relies on large language models (LLMs) to interpret and schedule user requests, I also designed for the complexities and unpredictability of AI output - ensuring users always had visibility and control over what the system was doing behind the scenes. My goal was to design and launch a robust MVP, then continue supporting the team post-launch with ongoing feature development and iterative improvements.
OUTCOME
Early Traction and Strong Early Engagement
The iOS app launched to early users and saw rapid adoption - within the first month, over 70% of users were active weekly, averaging 4 sessions per day and scheduling around 5–15 events per user each week
