Rule-based chatbots

We meet them first because they’re simple. A rule-based chatbot is a conversational system that lives on a script. She doesn’t improvise. Every path is written down somewhere, like notes on an index card.

Intents

Think of intents as her guesses about what we mean. If we type “book a flight,” the intent is “travel booking.” If we say “what’s the weather,” the intent is “forecast.” She doesn’t know everything, but she’s trained to look for those cues. Do this so users feel like she speaks their language.

Responses

Once she has the intent, she pulls up the right response. Sometimes it’s a short answer. Sometimes it’s a link. The point is consistency—so users get what they expect. We should keep responses short, too. Long answers feel like monologues, not conversations.

Decision Trees

The map under the hood is a decision tree. One branch for “yes,” another for “no.” If the user says something unexpected, the tree runs out of branches. That’s when she replies, “Sorry, I didn’t get that.” Good design means pruning dead ends and keeping the tree shallow. If users get lost, they quit.

Why It Matters

We may want fancy AI later, but rule-based chatbots remind us to start small. She is clear, predictable, and cheap to build. Users can get answers fast without knowing they’re inside a script. The trick is writing the rules as if we’re users ourselves.