Inspiration
Managing multiple calls and meetings often leads to lost details and missed action items. Writing follow-up emails after every discussion is time-consuming and repetitive. Transcripts from calls or Zoom meetings contain valuable insights that usually go unused. We wanted a smart assistant that could automatically extract the essence of every conversation.
What it does
Converts call or Zoom meeting transcripts into concise summaries. Identifies and highlights key discussion points and action items. Automatically drafts follow-up emails including next steps and important decisions. Generates a talk heatmap to visualize participation and engagement throughout the meeting. Provides both mobile and web interfaces — mobile for on-the-go email generation and web for deeper analysis and visualization. Saves users time by automating repetitive post-meeting tasks while improving collaboration and accountability.
How we built it
NeuraSeek and Gemini for AI insights and transcript generation Kotlin for android development NodeJs, React, tailwindCss for web development AuthO for authorization and login
Challenges we ran into
Integrating multiple AI APIs efficiently while maintaining low latency for real-time summarization and email generation. Fine-tuning the summarization and action extraction models to avoid generic outputs and capture meaningful insights. Balancing UI responsiveness with the computational demands of AI-powered features, especially on mobile devices.
Accomplishments that we're proud of
Successfully built a cross-platform solution integrating both mobile and web applications. Designed an automated email generator that produces clear, actionable follow-ups. Developed an interactive talk heatmap to visualize participant engagement during meetings. Seamlessly integrated with both Zoom and mobile call recordings, ensuring flexibility across use cases. Created a clean, intuitive UI/UX for effortless navigation and quick insight generation. Managed to automate post-meeting workflows that typically take hours — now achievable in minutes.
What we learned
The importance of designing scalable pipelines for handling large and unstructured transcript data The challenges of integrating audio transcription, summarization, and email generation into a unified workflow. How crucial UI/UX design is for simplifying complex features like talk heatmaps and summary visualization. The value of prompt engineering and fine-tuning models to extract meaningful action items from raw text. How collaboration across frontend, backend, and ML components can significantly improve overall efficiency. That even small automation steps can drastically reduce post-meeting workload and boost productivity.
What's next for Rize
Integrate real-time meeting analysis, enabling live summarization and action point detection as conversations happen. Expand multi-platform support, including iOS and desktop apps for a unified experience. Add email customization templates for different professional contexts (sales, project updates, client meetings, etc.). Enhance data privacy and encryption, ensuring transcripts and summaries remain fully secure. Fine-tune the AI models for domain-specific performance in areas like business meetings, customer support, and education.
Log in or sign up for Devpost to join the conversation.