Chatbot for Microsoft Price Calculator

The challenge

A technology-driven organization aimed to simplify Microsoft product pricing discovery by implementing an AI-powered chatbot capable of answering pricing-related queries across multiple Microsoft product lines. The objective was to build an interactive pricing assistant that helps users quickly retrieve accurate, contextual, and up-to-date pricing information without manually navigating lengthy pricing guides. The solution was designed to improve pricing transparency, reduce manual effort, and support multi-tenant access where customers can securely access only their relevant pricing information.

Key challenges

• Managing large volumes of pricing information across multiple Microsoft product categories
• Difficulty navigating chunked and text-based pricing documents manually
• Ensuring customer-specific pricing isolation in a shared environment
• Maintaining updated pricing information for accurate responses
• Supporting scalable and low-latency query processing for multiple users

Technology used

• Framework: LangChain for conversational orchestration and retrieval workflows
• Vector Database: Pinecone for semantic indexing and tenant-specific namespaces
• Embedding Models: OpenAI Embeddings for pricing document vectorization
• LLM Used: GPT-4o-mini for contextual answer generation and conversational intelligence
• Programming Language: Python for orchestration and backend implementation

What we built

• Data Preparation: Processes and structures Microsoft pricing guides for indexing
• Embedding & Indexing: Generates embeddings and stores tenant-specific pricing data in Pinecone namespaces
• Natural Language Querying: Accepts pricing-related questions through conversational input
• Semantic Retrieval: Retrieves top relevant pricing chunks based on user intent and context
• Response Generation: Uses GPT-4o-mini to synthesize retrieved pricing information into actionable answers
• Multi-Tenant Architecture: Ensures secure pricing visibility for each customer environment

Objectives

• Build an interactive chatbot for Microsoft product pricing assistance
• Enable natural language querying for detailed pricing information
• Improve accessibility to pricing guides across multiple Microsoft product lines
• Support multi-tenant data isolation for customer-specific pricing records
• Deliver fast and context-aware pricing recommendations
• Reduce manual navigation effort through intelligent search and retrieval

Outcomes

• Instant pricing Q&A delivering contextual pricing answers within seconds
• Improved accessibility to Microsoft pricing information across multiple products
• Secure multi-tenant isolation ensuring customer-specific data privacy
• Always up-to-date pricing retrieval through automatic indexing updates
• Scalable architecture capable of handling growing pricing data and user queries
• Enhanced user experience through conversational and actionable pricing insights