Convay Meeting Transcription

In fast-paced meetings, it’s easy to miss key points, and even easier to forget them later. Convay Meeting Transcription makes real-time conversations instantly actionable by generating accurate transcripts and AI-powered summaries. With the first-ever Bengali transcription support, it brings inclusive, professional-grade documentation to users across teams, languages, and platforms.

CATEGORY:

Web Design,

Product Design

ROLE:

UX Designer,

UI Designer,

UX Researcher

TOOLS:

Figma

Convay at a Glance

Convay is a video conferencing platform built for modern collaboration. Unlike traditional tools, it supports the full meeting lifecycle, from scheduling and hosting to AI-powered transcription, file storage, and post-meeting follow-ups.


Designed with scalability in mind, Convay now supports meetings with up to 10,000 participants and is trusted by governments and international organizations in over 46 countries.


Key features include:

  • High-quality video and audio conferencing

  • AI-based transcription and meeting summaries

  • Cloud storage for meeting files and chat logs

  • Real-time whiteboard, chat, and screen sharing

  • Enterprise-grade security with on-premise and cloud options


Convay brings everything into one platform to simplify meetings, improve productivity, and support high-stakes collaboration at scale.


Convay has been used in global events like SIDSSA 2025 and secured a €5M government contract through its scalable architecture and reliable UX.


Visit convay.com to learn more.

Feature Overview

Convay’s Meeting Transcription feature was designed to make meeting documentation effortless. Whether it's a team sync or a government-level discussion, users can record, transcribe, and generate AI-powered meeting summaries, in both English and Bengali.


With a simple workflow, users hit record, see real-time transcripts appear on screen, and generate structured meeting minutes with one click. The experience includes privacy-first consent, easy language toggling, and exportable files for follow-ups. The goal: eliminate the need for manual note-taking, while ensuring key decisions and action points are always captured, shared, and accessible.


This feature brought transcription to a wider audience, especially Bengali-speaking professionals, making Convay the first platform to offer Bengali AI summaries, and a competitive alternative to tools like Otter.ai or Zoom’s transcription.

Problem Statement

In fast-paced or high-stakes meetings, taking notes manually is distracting and error-prone. People either miss key points or end up spending more time organizing notes than participating in the discussion.


Existing transcription tools often cater only to English speakers, leaving Bengali-speaking users without accessible solutions. This gap creates friction for teams in Bangladesh and other regions where bilingual documentation is essential.


We needed a way to help users focus on the conversation while capturing everything accurately, in their preferred language, and make meeting takeaways instantly usable, not buried in recordings or forgotten notes.

My Role and Responsibilities

As a UX designer on this project, I was involved from the early research stage through to final handoff. My goal was to make meeting transcription and summarization accessible, especially for users who preferred Bengali or needed simplified documentation tools.


Working closely with developers, product leads, and another designer, I:

  • Mapped the end-to-end user flow, from initiating a transcription to exporting meeting summaries

  • Designed core screens for live transcription, AI-generated summaries, language toggling, and export

  • Tested two interface layouts for clarity and engagement, collecting feedback from internal stakeholders and beta users

  • Collaborated with the dev team to refine loading states, export logic, and multilingual switching

  • Improved the consent flow, ensuring users understood and agreed to recording before transcription began


Throughout the process, I focused on balancing ease of use with flexibility, ensuring that the feature could support quick documentation during fast-paced meetings — whether internal syncs or government sessions. I also maintained consistency with Convay’s broader design system, ensuring the mobile and desktop experiences felt aligned.

Design Process

Research & Insights
We started by interviewing users and internal stakeholders to understand what was missing in the current documentation process. Manual note-taking was slow, error-prone, and hard to share afterward. Existing transcription tools lacked support for local languages, especially Bengali, a barrier for many teams in government and internal settings.


We also reviewed tools like Otter.ai and Notta to identify gaps in multilingual support, exporting options, and visual clarity. Our focus areas became clear: real-time accuracy, language flexibility, and clear separation between transcript and summary.


“This makes it easier to share notes across teams. I don’t have to rewrite anything,” — said a participant during testing.



Ideation & Design
I sketched user flows that simplified the journey from audio input to polished, shareable output. Key design goals included:

  • Minimal effort from users before and after meetings.

  • Seamless export options (PDF, DOCX).

  • Ability to switch between Bengali and English instantly.

  • Clear distinction between raw transcript and polished summary.


To ensure the experience stayed lightweight on mobile, we prioritized layout clarity, readable type, and strong visual hierarchy, especially for bilingual content. Accessibility considerations included high contrast modes, legible font sizes, and feedback cues throughout the flow.


Iteration & Testing
We tested multiple versions of the summarization flow and UI labels to reduce friction and clarify intent. In one version, the summary button was embedded near the transcript; in another, it was shown separately. Users preferred the separate placement, it reduced confusion and helped them focus on one thing at a time.


Loading indicators were redesigned after feedback to better communicate AI processing. A multi-stage loader replaced the generic spinner, giving more transparency and reducing impatience during longer waits.


“At least I know something’s happening now,”

— one user said after seeing the updated loader.

UI Showcase

Transcription Page: Displays live transcriptions categorized by speakers.

Summarization Loader: Communicates progress during meeting minute generation.

Meeting Minutes Output: A professional summary with structured agenda and decisions.

Language Toggle: Allows users to switch between English and Bengali transcriptions seamlessly.

Translated Meeting Minutes: A professional summary with structured agenda and decisions.

Export Option: Provides users with a professional document of meeting minutes.

Confirmation Modal: Before starting transcription, users had to approve via a consent modal.

Challenges and Solutions

  • Challenge: Users felt unsure what would happen after clicking “Generate Meeting Minutes.”


    “I clicked it... but wasn’t sure if it was stuck or working.”

  • - one tester shared

    Solution: We added a clear loading screen and progress feedback to build trust during AI summarization. Microcopy like “Summarization in progress…” helped manage expectations and reduce abandonment.

  • Challenge: Confusion between transcript and summary outputs.
    Many users expected action items immediately, not realizing they were still viewing the full transcript.

    Solution: We introduced distinct tabs and icons to visually separate the two. This reduced misclicks and improved comprehension.

  • Challenge: Ensuring privacy and consent for transcription.
    Convay serves sensitive use cases, including government and institutional meetings. Starting transcription without clear consent could raise compliance issues.

    Solution: We implemented a mandatory confirmation modal with a short explanation and opt-out option before recording began, ensuring transparency and aligning with enterprise privacy standards.

  • Challenge: Making Bengali transcription feel reliable.
    Most users were familiar with English tools, but Bengali transcription was new territory. There was concern over accuracy, especially for formal discussions or accented speech.

    Solution: We worked with developers to test transcripts against various dialects. While the core model handled most inputs well, we also designed the UI to allow quick edits and speaker name changes.

    “It’s not perfect Bengali,” one user shared, “but it’s good enough that I don’t have to write anything from scratch.”


    That sense of effort reduction was a key validation point for us.


  • Challenge: Missed opportunity to revisit meeting takeaways.
    Some users wanted to bookmark summaries or access key decisions later, but the system lacked a lightweight way to organize past content.


    Solution: We proposed (and documented) a future enhancement: the ability to tag or favorite summaries. While not yet built, it aligned with upcoming Convay updates, adding product value through design insight.

Outcome and Impact

The transcription feature filled a long-standing gap for bilingual teams and government officials who previously had no reliable way to document meetings in Bengali. In internal testing:


  • We achieved around 90% accuracy in controlled environments

  • Meeting wrap-up time dropped by an estimated 60%, with summaries ready in minutes

  • All test users preferred the Bengali summary over manual notes, saying it made post-meeting follow-up feel “finally doable”


This early success positioned Convay as a credible alternative to platforms like Zoom or Otter.ai, especially in markets where language inclusivity and data trust are critical. While still in its early phase, the feature became a key part of Convay’s pitch to public institutions and local organizations.

Takeaways

When we started working on this feature, I thought it would be just another transcription tool. But as we tested it with real users, I saw how much it meant for people to have meetings documented in their own language. One tester smiled and said, “I’ve never seen Bangla meeting minutes look this polished.” That moment stuck with me.


Designing this made me rethink what accessibility truly means. It's not just about screen readers or font sizes. Sometimes, it's about seeing your own language represented, about being included without needing to ask for it.


It also pushed me to balance automation and trust. AI summarization sounds magical, but people still want control. So I had to learn how to make the interface feel smart but not distant, powerful but still human.


“The best features don’t just work,

they feel like they were made with you in mind.”