Chatbot: definition, limits and selection
Definition
A chatbot is a text-based dialogue system that takes questions in written language and answers them automatically. Simple variants work rule-based with fixed paths; modern variants use a language model and phrase answers freely. The chatbot is the surface of the conversation; answer quality depends on the knowledge and rule base behind it. It does not replace case work but takes recurring enquiries off people's desks.
Why the Term Matters
Chatbot is often used as an umbrella for any AI dialogue, which sets wrong expectations. Who separates the term cleanly distinguishes the pure text surface from the knowledge source and from voice-based systems. For mid-market firms this matters because a rule-based chatbot is cheap and predictable, while a free-phrasing one needs oversight and maintained content. The choice decides effort, risk and value.
When It Is Relevant
- The same questions repeatedly tie up time in service or reception.
- Enquiries arrive around the clock and answers should be instant.
- A first, well-bounded use case should demonstrate AI value.
- A form or FAQ page is rarely used and should become conversational.
When Not
When enquiries are complex, individual or emotionally sensitive, a chatbot frustrates more than it helps. For legally binding information an automatic text dialogue is risky too. And without a maintained knowledge base, a free-phrasing bot produces wrong answers that cost more than no answer at all.
Industries That Use It
- Online retail: answer standard shipping, return and availability questions without waiting.
- Housing management: capture tenant matters like damage reports or appointment requests in a structured way.
- Education providers: pre-clarify questions on courses, funding and registration conversationally.
- Municipal utilities: recurring citizen questions relieve the phone hotline.
Typical Use Cases
- FAQ assistant on the website referencing verified sources.
- Appointment pre-clarification that captures the matter and routes to the right team.
- Status enquiry on an order, job or case.
- Qualifying enquiries before handover to a human.
- Multilingual first-line information for international customers.
Decision Guide
| Signal | Chatbot fits | Alternative |
|---|---|---|
| Many identical text questions | rule-based chatbot | dynamic FAQ page |
| Free, open questions | AI chatbot with a knowledge base | company brain as the foundation |
| Binding legal information | rather not | a human with subject responsibility |
| Enquiries by phone | rather a voicebot | phone channel instead of text |
Example
An online retailer embeds a chatbot for shipping and return questions. In the potential analysis we define which questions the bot may answer, when it hands over to a human and where answers come from. Value stays measurable and risk stays bounded. This describes the method, not an asserted client metric.
Service Connection
Whether a chatbot is the right lever is clarified in the AI potential analysis; the suitable knowledge base is checked in the AI readiness check. For mid-market firms we bundle this in AI consulting for mid-market.
Sources
This framing describes common practice in deploying chatbots. Statements on data protection and transparency duties are general orientation and not legal advice; they must be checked case by case against the currently applicable legal texts.