Decision guide
Chatbot or custom workflow?
The terms chatbot, AI agent and automation are often mixed together. For a project decision, it matters more where the workflow starts, which action is allowed and who handles exceptions.
In brief
A website chatbot answers questions. A custom solution handles a defined workflow. Both can be useful, but they do not have the same scope or risk.
When a website chatbot is enough
A chatbot fits when the input is the website channel and the task is to answer recurring questions from approved knowledge sources.
Source ownership, a clear non-answer and a hand-off to a person belong in the product scope even though they are not spectacular features.
- Questions and answers are central.
- Knowledge is limited and approved by a subject owner.
- The system performs no unconfirmed actions in CRM, ERP or calendars.
When a workflow project is needed
Once emails or documents are classified, data is transferred, system entries are prepared or several exceptions need handling, a website chatbot alone is usually not the right project model.
This requires an owned workflow: input, review, action, exceptions, monitoring and rollback. Interfaces and data risks are assessed for the specific case.
Why the term AI agent decides little
A broad product name says nothing about which data is processed, which systems participate or whether an action may be carried out automatically.
Ask for a description of the normal path and its boundaries. Tests, responsibilities and costs can then be derived from it.
The smallest useful next step
If only channel, knowledge and support are open, use the plan finder. For a recurring company workflow, use the solution finder. If data class, exceptions or target process are still unclear, a process assessment is the more honest starting point.