For the creation of AI, as I have not yet had projects with this function, how would this type of real implementation of the tool work?

The IBM RPA deploys two machine learning mechanisms and an integration point: - Knowledge Base (KB) is a question-and-answer algorithm that learns from a database of questions and answers. Users upload a spreadsheet containing an FAQ (or any other set of questions and answers) and the system learns from them. Users can use this feature to ask questions in natural language and be able to answer them. This feature can be used by digital assistants (phone and chat) as well as any bot, since its input parameters are the text statement and the response is the set of possible answers, which can be used to respond to emails, tickets, etc. - Text Classification is a machine learning algorithm capable of classifying texts based on a set of examples. Our approach to text classification employs a proprietary algorithm that can generate a set of machine learning models over a set of examples and use these models in coordination to increase accuracy. - R. The IBM RPA has deep integration with R, providing the ability to use an R script in a bot. The integration takes care of translating R variables into IBM RPA variables and vice versa transparently. Documentation: https://www.ibm.com/docs/en/rpa/21.0?topic=automation-machine-learning