A knowledge manager is a person responsible for the bot training.

What is knowledge management?

When users ask Kbot to find solutions for various issues, the chatbot uses search engines and systems, such as RightAnswers, Google, Stack Overflow, etc. Then the chatbot returns articles that correspond to what the user is looking for. Kbot does not perform the search - the search systems do. Apparently, some results prove to be the perfect match, while some results might be less helpful. Users click either Like or Dislike in the chat, and that helps to create a general picture of appropriate and less appropriate results. The Expert Tools are designed to manage these results by accepting or rejecting users' feedback on articles.

Prerequisites

The Recommender System analyzes articles that have been liked by users. Based on this analysis, Kbot offers useful articles to users when appropriate, which speeds up the search for a solution.

Make sure that this system is enabled and configured.

1. Log in to Kbot under the administrator account.

2. Go to Content > Knowledge Bases.

3. Click Add in the upper left of the window.

4. Type the name of your new knowledge base and select Recommender System from the Engine drop-down list.

5. Fill in the fields.

6. Click Save.

Training

The Training view is designed to run sentences through all the defined knowledge bases. Review the obtained results and manage their position in the list of the most useful articles using the Like/Dislike mechanism.

1. Go to Content > Knowledge Training.

2. Select Training at the top of the window.

3. Click Add sentences to add sentences manually, upload a file, or use the data generated by reports.

4. Specify the knowledge base that Kbot will use to simulate user requests and search for answers and click Add. If no knowledge bases are selected, then all of them are used.

5. Select the sentence from the list to view the related articles.

6. Click Like if you think the article is useful.

Tutorial

• Learn how to train the bot.