Optimize your bot
Once the conversational data comes in which is a helpful source of information for improvement. This section helps you to optimize your bot based on real-life data.
In this 'Optimize your bot' section, you will learn how to:
- Refine your bot when you face issues in your bot.
- Improve the performance of your classes by using real user inputs.
- Read Teneo's build-in Suggestions when optimizing your own solution.
Refine your bot
At this point, we assume that you have already created several flows in your solution. As your bot grows, there are some things you should take into account that go beyond the tutorials you have accomplished so far. Let's take a step back and look at what you want to achieve with your bot, how you can proceed with testing and debugging, and what you can do in case your bot does not behave as expected.
Classes are an essential part of the backbone of your bot. By increasing their accuracy, you will get a better intent classifier model and thereby enhance your bot's ability to pick the correct answer for a certain question. In Teneo, you can use the conversational data generated by the users of your bot to optimize classes.
You can optimize your classes by either reviewing inputs that have been assigned to classes or by reviewing classes and balancing the amounts of training data for each class.
When building and maintaining a large solution, it can be a challenge to retain an overview of all the solution details. When multiple people are working on the same solution, it is easy to overlook something, to accidentally save an unfinished language condition, or to forget to follow best practices. To avoid these types of errors, Studio automatically scans your solution for common issues and reports them in a nicely categorized manner in the Suggestions panel.