OpenAI Enabled Knowledge Base Search: 4 Wk Proof of Concept

Tiger Analytics

Implement an OpenAI-enabled knowledge base search tool in just 4 weeks. This solution will help reduce the turn-around-time in responding back to a customer.

By allowing users to search using natural conversational language, OpenAI enabled search will enable operational efficiency and significantly reduce average handle time. The solution will be implemented using Open AI on Azure using tools like Blob storage, Azure Cognitive Search, Functions etc.

Potential Use Cases Include: • Contact center agents querying for information from policy documents. • IT help desk querying information from technical documentation, error notes etc. • Employee querying information for HR, reimbursement and medical insurance policy documents. • Call center agents querying information from adjuster notes, historical conversations, claim notes etc.

Typical Agenda:

  1. Week 1: Development Environment Setup & Configure Azure Services
  2. Week 2-3: Defining the logic within Cognitive services & OpenAI (Document Parsing, Retrieval, Summarizer)
  3. Week 4: Performance Evaluation and Iteration with the stakeholder

Deliverables:

  1. UI interface connected to the knowledge base (100 -200 PDFs) considered.
  2. Querying ability for this knowledge base based on natural language.
  3. Response generated as a summary and document from which the response was generated.

Scope is limited to a maximum of 200 pdf files

https://store-images.s-microsoft.com/image/apps.6519.6ab2136b-b792-4027-9122-f359dab73e59.2c2842e3-cb63-4f14-b88a-48cdb7f4c878.f0d26a3a-e110-40b1-a2a9-ad8179d37434
https://store-images.s-microsoft.com/image/apps.6519.6ab2136b-b792-4027-9122-f359dab73e59.2c2842e3-cb63-4f14-b88a-48cdb7f4c878.f0d26a3a-e110-40b1-a2a9-ad8179d37434