Skip to content

User Query Components

Every user message to a chatbot carries two kinds of information: what the user wants (the intent) and the specifics (the entities). This diagram takes the query "Book a flight to San Francisco next Tuesday" and shows how the system pulls the intent and entities apart, then recombines them into a single actionable request the backend can execute. Hover over any box for details.

Interactive Demo

Run MicroSim Fullscreen

You can embed this MicroSim in your own page with the following iframe:

1
<iframe src="main.html" width="100%" height="492" scrolling="no"></iframe>

Overview

The diagram has three conceptual layers:

  1. Natural language input (gray, top) — the raw query.
  2. Semantic understanding (middle) — the query branches into:
    • Intent (blue): "Book Flight", which selects the system action.
    • Entities (indigo), which contain typed values: Destination = San Francisco (orange) and Date = next Tuesday (green).
  3. Structured output (bottom) — the intent drives a System Action ("Search available flights"), while the entities fill Parameters (the dotted "fills slot" arrows). Together they produce the Actionable Request.

The key idea is that understanding a query means both classifying the intent and extracting the entities; neither alone is enough to act.

Lesson Plan

  • Separate the parts: Have students underline the intent and the entities in several example queries.
  • Trace the slots: Follow the dotted "fills slot" arrows and explain why the entities map to parameters.
  • Why both? Discuss what would go wrong if the system knew the intent but not the entities (or vice versa).
  • Write your own: Have students invent a query and draw its intent, entities, and the resulting actionable request.

References