EchoHer

Inspiration

Recently, a lot of us been involved in remote-based projects, whether that be internships, group projects, or other. We noticed how easy it is for certain voices to dominate and others to go unheard, especially in online settings. This inspired us to explore how technology could help foster more balanced and inclusive conversations.

What it does

EchoHer is an post-meeting AI assistant—a meeting analyzer. It analyzes who speaks, for how long, and how often they’re interrupted. It gives stats, flags imbalance, and sends a post-meeting report to help teams build more inclusive conversations.

Let’s say an intern was barely acknowledged in the call. EchoHer would highlight this in the report, with future suggestions—like prompting quieter voices or moderating over-talkers.

How we built it

We first brainstormed and discussed within ourselves, reading research and articles relevant to the issue of gender bias in AI, and potential solutions offered by AI. After we established an idea we all liked, we divided up roles within the group based on our specialties, splitting off to work individually in parallel while maintaining close communication.

We used a combination of React for the frontend, Python for the backend analysis, and natural language processing libraries to evaluate the transcripts. Transcription data was parsed and analyzed for metrics like speaking time, interruption frequency, and tone.

Challenges we ran into

The main challenge we faced was the lack of time - we had a fairly complex idea that we wanted to convey. We needed to strip down the idea into something doable within the time constraints, balancing the time complexity requires and showcasing our skills and ideas.

Accomplishments that we're proud of

We successfully implemented part of our product, including some backend analysis of speaker participation and interruption tracking. We also created a comprehensive design of the reporting dashboard. Most importantly, we transformed a conceptual social issue into a tangible technical solution.

What we learned

We learned a lot about balancing ambition with feasibility under pressure. We deepened our understanding of bias in communication, how it manifests in data, and how it can be measured and visualised. On the technical side, we learned how to process meeting transcripts using LLM technologies.

What's next for EchoHer

We have yet to fully implement our frontend design and link it with the backend. Beyond that, we want to expand EchoHer to provide live suggestions during meetings, improve interruption detection accuracy, and incorporate demographic insights (with consent) to further support equity. Eventually, we envision EchoHer integrating with popular platforms like Zoom, Teams, and Google Meet to become a seamless companion in creating fairer, more balanced discussions.