Skip to content
DataQI Documentation

Intent-driven chat

DataQI operates from user intent. Instead of forcing you to understand the internal agent architecture and manually select an agent, you simply talk to DataQI. The system automatically analyses your message and routes it to the most capable agent available.

How intent routing works

Section titled “How intent routing works”

DataQI evaluates each message against the pool of available agents and determines the best course of action:

  1. Implicit Routing: If your request is clear (e.g., “What is the standard tolerance for Model R-4000 rivets?”), DataQI routes the message directly to the appropriate knowledge agent.
  2. Conversation Continuity: If your next message is a follow-up (“And what about the shear strength?”), DataQI stays with the same agent, injecting your conversation history so the agent understands the context.
  3. Agent Switching: If your intent changes mid-conversation (e.g., “Now plot those tolerances on a chart”), DataQI seamlessly switches to an agent with charting capabilities. You receive a unified response, and the switch is handled transparently.

Handling ambiguity and errors

Section titled “Handling ambiguity and errors”

Unlike rigid chatbots that fail when a query doesn’t match an exact pattern, DataQI is designed to intelligently ask for help or clarification when it doesn’t understand your request:

  • Clarification: If your intent is ambiguous (e.g., “Analyse my data”), DataQI will pause and ask a clarifying question (e.g., “I’d be happy to help. Would you like a text summary or a chart?”) rather than guessing incorrectly.
  • Query Expansion: When your request is complex, DataQI doesn’t just rely on your exact wording. Internally, the platform generates several different ways to ask your question. This widens the search net, ensuring DataQI matches the most relevant source documents and tools even if you didn’t use the exact terminology found in the data.