An emerging EdTech startup was gaining traction with its online certification and upskilling courses, serving a wide audience of working professionals and college students. With more than 5000 learners active on the platform each day, the company relied on a Moodle-based Learning Management System (LMS) to deliver video content, assignments, and mentorship.
As the platform’s user base expanded rapidly, student queries became difficult to manage. Learners frequently submitted requests related to account issues, course access problems, technical glitches, and general inquiries about deadlines and certification.
A small support team of three members was handling an average of 143 tickets per day. Most queries came during weekends and course launch cycles, creating major bottlenecks in service delivery.
Some of the pain points included:
The company needed a scalable, intelligent support system that could work around the clock, deliver consistent answers, and reduce the load on human agents.
Edpilot Studio designed an AI-powered support chatbot using a Large Language Model (LLM) to automate the first line of student interaction. This approach was aimed at resolving the most frequent support issues instantly and seamlessly escalating complex queries to human agents.
Step 1: Data Mapping and Training
Our team reviewed and analyzed six months of historical support tickets. We extracted over 200 common questions and grouped them into intents and categories. This data was used to train the LLM to recognize student needs with high accuracy.
Step 2: System Integration
The chatbot was integrated into the existing tech stack, including:
Step 3: Multilingual Support
We configured the bot to support three languages: English, Malay, and Indonesian. This was crucial to serve the platform’s diverse student base across Southeast Asia. Custom language fallback rules ensured that if the bot did not understand a query, it could redirect the user politely to a human agent.
Step 4: Smart Escalation and Feedback Loop
A sentiment detection feature was added to identify frustrated users. If the bot detected dissatisfaction or repeated queries, it would escalate the conversation to a live agent. We also built a feedback loop to collect chat ratings and continuously retrain the bot based on unresolved interactions.
Within six weeks of deployment, the chatbot significantly improved the company’s support operations:
The deployment of a custom LLM-based chatbot empowered the EdTech platform to deliver consistent, real-time, and multilingual support to thousands of learners. It helped reduce operational load, improved student satisfaction, and introduced scalability into a previously stretched support function.
This project demonstrates the real-world impact of integrating conversational AI into EdTech workflows. By automating repetitive support tasks, the company increased efficiency and enhanced the learning experience without compromising on quality.
5,000+ daily learners
143+ support tickets daily, long wait times
LLM-powered multilingual chatbot with Moodle & Zendesk integration
65% fewer tickets, 3x faster resolutions, 92% satisfaction
Your partner in pioneering AI-driven solutions, engaging elearning experiences, and future-ready applications to elevate your digital strategy.
Edpilot Studio Pvt. Ltd.
Street No. 82, Newtown, Kolkata 700165, IN