Glossary
Agentic Chatbot vs. Rule-Based Chatbot
A rule-based chatbot follows a fixed decision tree: it matches what a user types to a pre-written reply, and can only handle the exact paths its designer scripted. An agentic chatbot instead reasons about the request, decides what steps are needed, and can handle questions its designer never explicitly anticipated.
Rule-based chatbots are the oldest and simplest kind of chatbot: a designer builds a decision tree ("if the user says X, reply with Y"), and the bot follows it exactly. They’re predictable and easy to test, but brittle — a question phrased slightly differently than expected falls outside the script and the bot fails.
Agentic chatbots use a language model to understand intent rather than match exact phrasing, and can reason across multiple steps — checking a condition, looking something up, then deciding what to say — instead of following one fixed path. This makes them more flexible, at the cost of being less perfectly predictable than a rigid script.
In practice, most real chatbot platforms sit somewhere between the two: a designed conversation structure (so the business controls what topics the bot handles) combined with an AI layer that handles the actual language understanding and reasoning within that structure. That’s the model Basium uses — you design the flow, the AI handles understanding and multi-step reasoning inside it.