AI meeting notes
As part of the internal product AIDEA, a module was developed for meeting recording with transcription, AI-powered summarization, and a built-in AI assistant
Goals
🎯
Make meeting outcomes immediately useful (summary + structure)
Reduce time spent on processing information
Simplify the meeting recording process
Increase the value of the AIDEA platform
Pain points
🚧
- Users spend a lot of time reviewing and going through recordings
- There is no unified tool for recording and analyzing meetings within the product
- Important agreements get lost or forgotten
Workflow
🚀
- Research
- Received a task to create a service for recording offline meetings. There were no specific requirements
- Spoke with the stakeholder to gather initial information
- Collected references and studied similar solutions on the market
- Analyzed meeting workflow scenarios
- Identified core user needs
- Defined the key value — quick access to the essence of a meeting
2.Prototyping
- Initially we wanted to build a standalone service that would not overlap with our flagship product AIDEA
- Built the first clickable prototypes in Claude
- Conducted a review with the team and received approval
3. Solution reassembly
- After that, the security team stepped in and said the service must be built within AIDEA
- Adapted the scenarios to the existing system. As a result, the user can record a meeting, which is then processed, transcribed, and an AI meeting summary is generated. The user can also download the summary and transcription
4.Design
- Designed the final mockups
- Worked through key states
- Prepared the designs for developer handoff
5.Developer handoff
- Handed off designs to development with full documentation
- Conducted a design review of the finished solution
- Released to production
6. Metrics analysis
- Reviewed the metrics
- Surveyed users on the service's usability
- Collected feedback for future improvements
Workflow
🚀
- Research
- Received a task to create a service for recording offline meetings. There were no specific requirements
- Spoke with the stakeholder to gather initial information
- Collected references and studied similar solutions on the market
- Analyzed meeting workflow scenarios
- Identified core user needs
- Defined the key value — quick access to the essence of a meeting
2.Prototyping
- Initially we wanted to build a standalone service that would not overlap with our flagship product AIDEA
- Built the first clickable prototypes in Claude
- Conducted a review with the team and received approval
3. Solution reassembly
- After that, the security team stepped in and said the service must be built within AIDEA
- Adapted the scenarios to the existing system. As a result, the user can record a meeting, which is then processed, transcribed, and an AI meeting summary is generated. The user can also download the summary and transcription
4.Design
- Designed the final mockups
- Worked through key states
- Prepared the designs for developer handoff
5.Developer handoff
- Handed off designs to development with full documentation
- Conducted a design review of the finished solution
- Released to production
6. Metrics analysis
- Reviewed the metrics
- Surveyed users on the service's usability
- Collected feedback for future improvements
Results
🏆
1 month after the feature launch
- 46% of users interact with the AI assistant
- Meeting analysis time reduced: from 20–30 minutes down to 5–7 minutes
- +27% increase in time spent in the product
Results
Managed to transform the offline meeting recording process into a full-fledged information management tool: users get to the essence of discussions faster, lose track of agreements less often, and return to recordings more frequently. The feature enhanced everyday scenarios within the product and made AI a useful part of the workflow rather than a standalone function
AI assistant interaction
46%
Meeting analysis time
-70%
Time in product
+27%
