Transform RPA with AI and ML

ML in Robotic Process Automation

• • Knowledge base
    
    • • In IBM RPA, you can create knowledge bases for your robots to answer questions by searching answers in a knowledge base.The knowledge base is a machine learning model that stores data in the form of questions and answers.
      • We can use the RPA tools Machine Learning Model builder to select the knowledgebaseV2 algorithm and provide the knowledgebase file.
      • This knowledgebase file is made up of text organised into three tables, each in its own spreadsheet.The inner structure of the file must include three spreadsheets named KB, Word, and Synonyms.
      • Insert the knowledge base's set of questions and answers, context label, and tags associated with the question into the KB spreadsheet.
      • Once the Knowledgebase has been successfully created, we can select the newly created kb and look for sample questions and answers that have been trained.During this phase, we can also train the KB with additional related questions and answers.
      • When the training is finished, publish it.We can now begin using published KB directly from the RPA bot script, using the AnswerQuestion function under NLP.

• • Text classification
    
    • • IBM RPA allows you to create text classification models for document labelling.
      • The Text Classifier model represents a collection of text documents organised and classified in tagged directories.As a result, each text document is assigned one tag.To train the model, the text classifier algorithm combines several algorithms.
      • We can choose Train Classifier function from RPA tools Machine Learning and provide folder path of tagged directories.
      • We can now use the published model directly from the RPA bot script, using the ClassifyText function under NLP.

• Increase productivity  

• Reduce costs

• Increase accuracy

• Enrich the customer experience

In the asset management industry, for example, bots were used to provide first-line customer support and pricing quotations. This optimization significantly reduced the time it took to respond to customer questions, thus improving the customer experience and streamlining the purchasing process.