Financial FAQ and Knowledge Assistant Agent

Agent Description:

The Financial FAQ and Knowledge Assistant Agent is an intelligent assistant designed to provide fast, accurate answers to customer or employee queries related to financial products and company policies. It leverages internal FAQs and product documentation to deliver concise, context-aware responses.

Purpose and Components
  • Purpose: To assist users with instant and accurate answers to financial product and service questions by querying structured internal knowledge sources.
  • Components:
    • A knowledge retriever to ingest and structure content from FAQs and documents.
    • An answer generator that crafts concise, user-friendly responses using retrieved data.
Supported Capabilities
  • Reading and indexing internal FAQ PDFs and policy documentation.
  • Structuring extracted content into searchable formats for efficient querying.
  • Understanding natural language questions related to finance, billing, refunds, plans, troubleshooting, etc.
  • Generating context-aware, precise answers aligned with internal documentation.
  • Supporting live FAQ/chatbot experiences for internal or customer-facing use.
LLM Used
  • Google Vertex
    Note: To learn more about the LLM and to modify its behavior, refer to the Configuring LLM settings section.

Sub-Agents

1. Knowledge Retrieval Agent

  • Role: Knowledge Retriever.
  • Scope: Connects to the internal knowledge base and prepares structured FAQ data.
  • Description: This subagent reads uploaded PDF or text-based FAQ and policy documents, identifying and structuring relevant sections. It covers topics such as billing, subscription plans, refunds, sign-up processes, troubleshooting steps, and integrations. The extracted data is organized into a searchable format that supports downstream question-answering tasks.
  • LLM Used: Google Vertex (inherits from parent).
  • Tools used: MCP Client Tool

2. Financial FAQ Agent

  • Role: Answer Generator.
  • Scope: Generate clear, accurate responses to product and finance-related questions.
  • Description: This subagent receives user queries and uses data retrieved by the Knowledge Retrieval Agent to construct concise, context-rich answers. It ensures that the response strictly adheres to information from the internal knowledge base. It prioritizes clarity, correctness, and alignment with the company’s policies and product details.
  • LLM Used: Google Vertex (inherits from parent).
  • Tools used: MCP Client Tool
Tools Used

MCP Client Tool for document ingestion and preprocessing for RAG Internal Knowledge Base Search

Note: For details on modifying the Tools, refer Tools Library section.
Agent Workflow Behavior Summary
  • The Knowledge Retrieval Agent processes uploaded FAQ and policy documents, organizing the information into structured formats.
  • The Financial FAQ Agent handles user queries by leveraging the structured knowledge to deliver concise, contextually accurate answers based on internal documentation.
Sample Questions:
  • How do I reset my ConnectSphere password?
  • Do you offer refunds or discounts for educational institutions?
  • Does ConnectSphere integrate with Google Drive or Slack?
  • What is the file storage limit on the Basic plan?