Chatbot – what does it mean?

How it works

A question typed into the window at the bottom of a website can be answered by two kinds of program. The rule-based version responds to a keyword with a pre-written text. The AI-based version interprets the question and writes an answer from the company’s own content: website, product pages, price list, documents.

If there is no content in the knowledge base, the assistant does not guess: it asks for the contact details and hands the question over to a colleague. The source used is shown next to the answer.

An example at a company

“When will my parcel arrive?” and “How long does delivery take?” are two separate cases for a rule-based program, and one for an AI assistant. The AI assistant answers from the delivery terms page and also shows the source.

With a rule-based solution, a separate answer has to be written for every new way of phrasing a question. If the delivery deadline changes, the text has to be rewritten by hand.

What this means for business leaders

This is the term that has caught on in the market, but the price and risk of the two types differ. Ask the provider which one you will get, and whether the answer comes from the company’s own content.

If the knowledge base is not updated, the assistant may give outdated information, so ask who updates the content. The risk: the customer is given the wrong price or deadline.

How we do it

Onboarding is done by our team, and before go-live we check the assistant’s answers with ten test questions. For installation, access to edit the website is enough, or your web developer places the code snippet.

The assistant answers in the customer’s language, even if the knowledge base is in another language. The AI disclosure appears in the header and in the first message.