AI-powered University Student Support Chatbot Using AWS Serverless Technologies and Claude AI

Abstract

Artificial Intelligence has boosted the adoption of chatbot systems across higher education, yet many current chatbots still face significant limitations. Existing systems provide shallow or generic responses, struggle with informal or vague student language, fail to maintain context across multi-turn conversations, and function as FAQ tools without integration into real university services. Many also lack scalability and the ability to support multiple administrative domains within a single platform. These flaws reduce reliability and limit their usefulness in real student environments where inquiries are diverse, urgent, and most of the time complex.

To address these challenges instantly, this research aims to design, develop, and evaluate a cloud-native conversational chatbot built on an architecture designed for real-time intelligence. A web frontend routes student messages through API Gateway into Amazon Lex V2, where intents are recognized before being processed by AWS Lambda. Lambda acts as the system’s logic, retrieving data from DynamoDB for students’ records, fees, and appointments while optimizing Claude Sonnet 4.5 for deeper semantic understanding. Additional integrations like Amazon SES and external APIs extend the chatbot’s functionality, making the system adaptive and responsive to support across the university’s digital ecosystem.

The chatbot supports more than ten major university service areas, including tuition inquiries, academic advising scheduling, visa guidance, dormitory information, GPA Calculation, mental health resources, health and car insurance policies, and part-time job guidance. Functional evaluation was conducted using 63 test cases covering simple queries, informal language, typing errors, vague requests, and multi-turn conversations. Testing results showed strong intent recognition, accurate data retrieval, effective handling of informal inputs, and reliable multi-turn conversation performance.

Although future improvement is needed in multilingual support and handling of highly unstructured messages, the prototype illustrates a compelling direction for higher education. It shows how intelligent conversation delivered instantly, clearly, and empathetically can transform the student support experience, reduce administrative strain, and bring institutions closer to the students they serve.

Keywords: Chatbot; Amazon Lex V2; Artificial Intelligence; Higher Education