Designing AI-native meeting recorder.
As the founding designer at Clarity, I led the end-to-end design of our AI-native meeting recorder from 0 to 1. The journey was highly iterative—filled with experimentation, scrapped concepts, and continuous learning. I shaped the product experience across multiple workflows, built and scaled our design system, aligned cross-functional teams on key decisions, and actively promoted a culture of design thinking. I also worked closely with product data to uncover new opportunities and refine our direction.
So much of conversational insights are lost, leading to misalignment, and repeating conversations.
In today’s fast-paced work environment, meetings are often filled with crucial insights, decisions, and action items. However, much of this information gets lost due to the limitations of manual note-taking or poor record-keeping. As a result, teams struggle to stay aligned, and important details are forgotten or misunderstood. This leads to unnecessary repetition of conversations and delays in decision-making.
How might we help people leverage their meetings' data to boost productivity?
Conversational data is one of the most valuable resources for teams, and AI allows us to unlock its full potential to drive productivity. We focused on capturing meeting data, as we believe it holds significant market potential for delivering enhanced features and a superior user experience.
Supporting people to record every important meeting, anytime and anyhow.
Clarity bot can be set to automatically join scheduled meetings or ad-hoc meeting. It will record meeting video and audio data.
A go to meetings section in the app combining recorded and upcoming meetings Main space shows all the recorded meetings. Right panel to see upcoming calendar meetings - allows to modify meeting bot preference, recording controls.
Start a new meeting in the right panel by adding meeting link any time.
After deploying the app to pilot customers, we got some good feedback and learnings that helped us improve.
We added both bot and audio recording preferences to give people freedom to choose recording type based on their needs.
We did not remove bot as we thought video recording of a meeting holds significant value.
We replaced sidebar of upcoming meetings to a calendar view to fit the mental model better. The top of calendar had recording panel to see updates and controls related to all recordings.
Calendar allows to change recording preference to assistant/bot or audio beforehand, so that Clarity can remind you at the right time.
Once the recording has started, all recording related controls also are shown in the same panel to have a constant frame of reference.
We also provided a faster way to audio record on your desktop. Using a shortcut command, users can access the recording toolbar and control the recording.
After deploying the app to pilot customers, we got some good feedback and learnings that helped us improve. At this stage, we also employed Full story to analyze user sessions to get better insights.
To provide a clear positioning, we tried to position the meeting recorder as a personal productivity tool that helps you capture every insight through audio recording every conversation.
Creating a focused experience to support users record and control their meeting schedule for the day.
Providing one primary way to start and stop the recording.
After deploying the app to pilot customers, we finally received positive feedback appreciating the improved UX.
Integrating traditional manual note-taking.
We also tried to integrate manual note-taking into the flows at different degrees to see which one fits the workflow best.
After deploying the app to pilot customers, we got some good feedback and learnings that helped discover note-taking experience.
After introducing audio recording, we integrated floating note panel into the audio recording toolbar to keep the experience non-distracting and in-context.
After deploying the app to pilot customers, we got some good feedback and learnings that helped us improve. At this stage, we also employed Full story to analyze user sessions to get better insights.
Next iteration introduced only one notepad in the meeting page on the app, and we designed app to work better in minimzed mode during meetings as a notepad to keep the experience consistent.
Maximizing meeting insights for maximizing productivity.
We improved our meeting insights gradually based both on customer feedback and product direction.
Based on our initial direction to support customer feedback, we had editable summary and product feedback as main meeting insights.
After deploying the app to pilot customers, we got some good feedback and learnings that helped us improve.
After deploying the app to pilot customers, we got some good feedback and learnings that helped us improve. We also got insights from Fullstory user sessions.
After all the feedback, we decided to cut down on visual complexity to make the experience really focused and minimal with only essential insights.
Templates to modify the summary based on your needs. Icons were geenrated using Midjourney.
Surfaces to ask questions across meetings for faster insights.
Sharing and collaborating on meetings for team productivity.
Meetings are essential data sources for team workflows. Based on our pilot users workflows, we had to support sharing and collaboration to integrate better.
We added channels to share meetings with relevant team without sacrificing privacy.
Launching self-serve onboarding to scale.
After manually onboarding every potential user, we decided to launch a self-serve onboarding that can scale to more users.
I created a marketing website on Webflow to start onboarding new users.
Onboarding flow from the marketing website till home page.
Settings for user controls.
As we moved more users to self-serve, it was important to have settings for independent app controls and customizations.
Getting feedback early.
In high-speed startup environments, open and honest feedback—no matter how uncomfortable—is key to staying efficient and aligned.
Having a core workflow is important.
In all the iterations we did for product, one thing became clear that having a core workflow to sell and make users adapt is really important. Feature bloat to shine helps no one.
Adding AI features needs a balance between user needs and AI capabilities.
Whenever thinking of adding AI features to the product, only focusing on user needs and their wants might not lead to innovation. AI is such a new technology that it's hard for users to even imagine what's possible. IT requires us designers carefully think how AI can innovate in the areas no one has thought of before.
