1 month
Project timeline
Active Lens AI
Redesigning an AI meeting assistant from passive summaries to real-time intelligence.
Active Lens is a desktop AI meeting assistant that records and transcribes calls, generates notes, and turns conversations into actionable next steps. I redesigned the existing product experience and explored how AI could become useful during the conversation, and not only after it.
What I did
I worked across the core product experience, from fixing foundational UX issues to designing new AI-powered interactions.
- Audited the existing desktop experience and identified key UX issues;
- Restructured navigation and simplified core product flows;
- Redesigned the desktop application and established consistent interaction patterns;
- Designed new real-time AI experiences for live meetings;
- Explored participant recognition, contextual intelligence, and AI-powered actions;
- Built a scalable design system for existing and future product features.
1 Product Designer
End-to-end design ownership
2-3 core flows
Substantially redesigned
4 AI capabilities
Designed for live meetings
From understanding the existing product to shipping the redesign
I started by understanding how the product worked today, then moved from identifying friction to restructuring the experience, exploring new AI capabilities, and refining the solution with the team.
Discovery & Audit
I started with stakeholder interviews, an audit of the existing desktop app, and a review of the main product flows. The goal was to understand the underlying UX problems before changing the interface.
Confusing navigation
Some parts of the product were difficult to discover or accessible only through indirect actions.
Inconsistent behavior
Popups and full-window experiences were mixed without a predictable pattern.
Unclear hierarchy
Naming, structure, and transitions made orientation harder than necessary.
Unnecessary friction
Key actions depended on other screens or required avoidable steps.
Competitive Research
I reviewed competing AI meeting tools to understand how they approached live transcription, meeting intelligence, AI assistance, and post-call workflows.
The research helped identify established patterns, gaps in the live-meeting experience, and opportunities for Active Lens to provide useful intelligence while the conversation was still happening.
User Jobs & Experience Map
Instead of treating individual screens as separate experiences, I mapped the product around three key moments and the main job users need to accomplish in each one.
Before the meeting
Prepare quickly and access relevant participant or client context.
During the meeting
Stay focused while capturing information and getting useful assistance.
After the meeting
Review what happened and turn it into clear next actions.
Design Principles
Principles that guided the redesign.
Make navigation predictable
Users should always understand where they are and how to reach key areas.
Keep AI contextual
Surface information when it becomes relevant instead of making users search for it.
Reduce distraction during calls
AI should support the conversation, not compete with it.
Keep users in control
AI can suggest and automate, but important actions should remain understandable and reviewable.
Key Product Decisions
The redesign focused on fixing the core UX first, then using that stronger foundation to introduce new real-time AI capabilities.
Rebuilding the navigation
I rebuilt the navigation around a clearer hierarchy, removed hidden paths, and standardized how users moved between screens. Key destinations became directly accessible, screen behavior became more predictable, and dependencies between different parts of the product were reduced.
Bringing AI into the live call
Active Lens was already useful after meetings. The bigger opportunity was making that intelligence available while the conversation was happening.
- Voice-controlled AI assistance
- Participant recognition using CRM and web context
- Entity detection for companies, events, and dates
- Real-time action suggestions
Turning context into action
The goal was not only to surface information, but to make it useful. Active Lens could recognize important moments during a conversation and turn them into actions users could review - from updating CRM information and creating tasks to scheduling follow-ups or working with AI-generated content.
Design System
Alongside the redesign, I created a new design system to bring consistency to the existing product and provide a scalable foundation for new AI functionality.
- Reusable components
- Consistent interaction patterns
- Clear states and behaviors
- Light and dark interfaces
- Shared visual rules across the product
This gave design and development a common system for evolving the product without introducing new inconsistencies.
Validation & Iteration
I reviewed the main flows and design decisions with the founder and Design Lead throughout the project. Their feedback helped refine product logic, interaction details, and how the new AI functionality fit into the existing experience.
I also worked closely with development to make sure the proposed interactions were technically feasible before final handoff. The redesigned experience was then implemented and shipped to production.
Final Experience & Outcome
The final product connected the experience across three key moments:
- Before - prepare and access relevant context
- During - capture the conversation and get real-time AI assistance
- After - turn the meeting into notes, decisions, and actions
Clearer structure
More predictable navigation and screen behavior
Simpler workflows
Core product flows restructured and streamlined
Scalable foundation
New design system built from scratch
Shipped product
Redesign and new AI capabilities released to production