Sonic Support Agent - Case Study


In the rapidly evolving landscape of decentralized finance (DeFi), Sonic DEX stands out the most popular multichain decentralized exchange built on the Internet Computer Protocol (ICP). Offering a broad spectrum of DeFi services, including token trading, liquidity provisioning, swaps etc. Sonic DEX has quickly gained a substantial user base in the ICP Ecosystem. But, with growth comes the challenge of scaling customer support to meet the increasing demands of users. This case study explores how Sonic DEX addressed this challenge by integrating custom-trained AI Agent to change how their customer support system operated.


As of 2023, Sonic DEX experienced rapid growth, with its community expanding to hundreds of thousands of users and its trading volume surpassing with an daily average trading volume between 300k to 500k. With a relatively small team dedicated to customer experience, the platform faced the though task of managing an increasing volume of user inquiries. Daily, the support team encountered over 40 new tickets, alongside active engagement in their Discord server with over 18,000 plus members and significant traffic to their DEX. The primary challenge was to maintain high-quality, personalized support without compromising on response time or accuracy.


To tackle this challenge, Sonic DEX partnered with ELNA, a leader in AI solutions, to deploy custom-trained AI Agent designed to enhance the user support experience. The AI Agent were trained on a comprehensive knowledge base that included Sonic's operational protocols, DeFi concepts, and the specific nuances of the Sonic DEX platform. This training ensured the AI could handle a wide array of user queries, from basic informational requests to more complex technical issues.

The integration of these AI agents into Sonic's support ecosystem was a strategic move to provide real-time, efficient, and consistent responses to user inquiries. The AI system was designed with clear guardrails to ensure that communications remained within the bounds of Sonic's brand voice and compliance requirements. Additionally, the AI was equipped to learn from each interaction, continuously improving its responses over time.


The implementation process involved several key steps:

  • Training the AI: Leveraging ELNA's platform, Sonic's team meticulously trained the AI on a curated dataset that included FAQs, user manuals, and historical support tickets.

  • Integration: The AI was seamlessly integrated into Sonic's the support section of their website.

  • Monitoring and Feedback: Sonic's team established a monitoring system to oversee the AI's performance, collecting user feedback to identify areas for improvement.


The deployment of AI-driven customer support had a very transformational impact on Sonic DEX's operations:

  • Improved Response Time: The average response time to user inquiries decreased significantly, with the AI agent handling a large volume of queries instantaneously.

  • Increased User Satisfaction: Users reported higher satisfaction levels due to the timely and accurate support, enhancing their overall experience on the platform.

  • Scalability: The AI system's ability to handle multiple inquiries simultaneously allowed Sonic's human support team to focus on more complex issues, thereby improving the efficiency of the support function.

  • Continuous Improvement: The AI's learning mechanism ensured that the quality of support improved over time, with the system adapting to new user queries and platform updates.


The integration of a custom-trained AI agent into Sonic DEX's customer support system represents a significant milestone in the platform's commitment to providing exceptional user support. By leveraging AI, Sonic DEX has not only enhanced its operational efficiency but also set a new standard for customer service in the DeFi space. This case study underscores the potential of AI to transform customer support, offering insights that could inspire other platforms within the DeFi ecosystem and beyond.

Ref : 4

Last updated