Media
Harvard TutorBot: Proof of Concept for Developing Cutting-Edge AI Tools
10+ months advancing AI innovation in education.
Results
Developed an AI-powered chatbot (TutorBot) that provides 24/7 access to Harvard’s academic resources.
Engineered a multi-format data ingestion pipeline enabling intelligent search and context-aware responses.
Delivered a Microsoft Copilot-based agent that preserved content attribution and source credibility.
Strengthened Unosquare’s role as Harvard’s strategic AI innovation partner for future EdTech initiatives.
Highlights / Intro
Harvard Business School sought to unlock continuous learning for students by leveraging its vast library of academic content through an AI-driven assistant.
The goal: to make world-class resources available on demand while maintaining data integrity and reliability.
unosquare partnered with Harvard to design and deliver a proof-of-concept chatbot; TutorBot powered by Microsoft Copilot and AWS OpenSearch, capable of understanding academic context, referencing credible sources, and providing personalized assistance to learners.
Challenge
Harvard needed a way to:
Transform thousands of unstructured academic documents into AI-readable formats.
Ensure every AI-generated answer included source attribution and contextual accuracy.
Create a scalable data pipeline capable of handling complex academic queries in real time.
Solution
unosquare engineered a comprehensive AWS-based ingestion pipeline that processes multi-format academic content into structured data ready for Microsoft Copilot.
By combining custom context layers with OpenSearch and LLM services, the chatbot delivers real-time, citation-based answers aligned with Harvard’s quality standards.
The MVP demonstrated how AI could enhance student engagement, reduce research time, and open new monetization opportunities through licensing and content access models.
Today, unosquare represents 40% of HBP’s engineering organization, providing ongoing product development, modernization, and cloud support across divisions.
Scope:
- Data Engineering & AI Integration
- Microsoft Copilot Customization
- End-to-End PoC Development & QA
Tech Stack
AWS (OpenSearch, LLM services), Microsoft Copilot, Custom Ingestion Pipeline
Centers of Excellence Involved
AI COE → Built and fine-tuned the LLM integration and context-aware retrieval system.
Data COE → Designed ingestion and indexing pipelines for multi-format content.
Product Engineering COE → Delivered end-to-end MVP architecture and UI integration.
QA COE→ Validated AI accuracy and response attribution compliance.
AI meets academia, bringing Harvard’s resources to every student, anytime.
Next Starts Here
