In conversation with Anik Das, Director of Engineering at Yellow.ai
We raced to build video chat in-house with an open-source WebRTC solution. It took us 3 months to build an MVP but polishing it for customers took a brutal 4-5 more. Calls that happened between countries suffered from high latency, and ensuring the reliability of every call was a challenge! We needed a solution with rock-solid stability, seamless integration with our existing systems, readily available native SDKs for mobile and web development, and proper support
Yellow.ai helps enterprises deliver omnichannel customer support through email, live messaging, and SMS. Recognizing the growing need for personal connections, especially after the COVID-19 pandemic, Yellow.ai embraced live video as the future of customer experience.
This shift aligns with industry trends, with live video support usage growing by as much as 233%. This allows enterprises to offer their clients a more engaging experience.
Yellow.ai took an ambitious route by building a custom live video solution within their platform using Janus (an open-source WebRTC server). This approach offered flexibility but also presented challenges. Developing a solution that could seamlessly scale to handle enterprise-level video interactions required extensive time, resource allocation, and deep engineering expertise.
Here are the challenges Yellow.ai had to deal with:
It took 3 months to build a MVP version, then another 4-5 months to get the live solution ready for enterprise customers. This delay hurt their ability to deliver as per their roadmap and meet customer timelines.
Janus offered a starting point for them to build live video, but Yellow.ai's engineers faced a steep learning curve with the solution. The open-source project's limited documentation and small developer community hampered support.
High latency in cross-continent calls revealed the drawbacks of a centralized server. While a decentralized model was the answer, it had its own complexities: managing servers, ensuring scalability, and maintaining the infrastructure.
Maintaining the live video feature post-launch became a resource drain. Constant bug fixes, performance optimization, and edge case management diverted engineering focus from Yellow.ai's core mission: automating customer support interactions.
Hit with roadblocks, Yellow.ai recognized the need for a sleeker, more scalable live video solution.
Enter Dyte.
Opting for Dyte provided us with the seamless integration and robust support we needed to enhance our platform's live video capabilities efficiently, marking a significant step forward in our mission to deliver exceptional customer support.
Yellow.ai prioritized enhancing its product with live video without sacrificing its core text and voice automation focus. Partnering with Dyte aligned perfectly with their strategic goals and operational needs. Key factors included a quick launch timeline, seamless scaling for enterprise use, and customization options.
Here's why Yellow.ai opted for Dyte:
Yellow.ai was impressed by Dyte's effortless integration process. They were able to add live video features to their platform within 2 weeks that saved them a lengthy development cycle.
Yellow.ai knew their customers wouldn't wait for support, which is why Dyte's proactive support guarantee was a game-changer. This ensures any issues are resolved quickly, minimizing downtime and keeping users happy.
Yellow.ai's education and healthcare clients have strict record-keeping requirements. Dyte's support for webhooks and a call recording pipeline addressed these needs perfectly. Webhooks allow for real-time notifications about call events, enabling automated workflows and data capture. The recording pipeline ensures secure storage of calls, compliance with regulations and facilitating future reference.
Team at Yellow.ai could use Dyte's UI kit to tailor the video call interface to perfectly match their brand aesthetic, while maintaining a smooth and responsive user experience across devices.
Meeting regional video recording requirements is crucial for Yellow.ai's clients, particularly those in finance and pharma. Dyte's ability to deliver this feature gives Yellow.ai confidence that its customers can adhere to strict compliance standards.
Our users enjoy a flawless live video experience thanks to Dyte's powerful streaming tech and customizable UI kits. Their robust infrastructure cuts down on lag, keeping our global calls smooth and responsive
Yellow.ai is doubling down on live interactions in its platform. The company plans to add background blur and plugins to its platform. Background blur would make it easier for support agents to take the conversation from anywhere without worrying about their background. Plugins, on the other hand would offer customers more ways to explain their problem / support reps to help customers solve their challenges.
We’re the best Live Video and Voice SDK, see for yourself
“We wanted to bring a video based interaction with users on our platform. Dyte checked all the right boxes for us!”