Introduction to Robotic Process Automation (RPA)

Why is this important?

Testing a bot for errors is crucial, particularly when it involves human interaction. This is because the interactions between a bot and a human can be complex and nuanced, and a bot that is not thoroughly tested can result in negative user experiences. For instance, if a bot provides inaccurate responses or fails to understand a user's queries, it can lead to frustration and confusion on the user's end. In addition, a bot that is not well-tested may have security vulnerabilities that can compromise sensitive user information. Therefore, rigorous testing is essential to ensure that the bot functions as intended and meets the expectations of users. This testing should include both functional and non-functional testing, including testing the bot's responses, its ability to handle unexpected inputs, and its performance under different loads. Ultimately, thorough testing can help ensure that the bot provides a positive and seamless experience for its users.

What will we do?

In this class we undertake our first practical application of building an automation using a guessing game bot as an example. We will first get the base model running and then investigate the problems with poor quality data inputs. We will quickly see the problems from poor data inputs in this model, as the bot will crash and cease to function. We will then move towards building out the controls that will help the bot flag poor data inputs that serve as a starting point to considering control systems for automation solutions.

Review and Extension:

Last class, we discussed the importance of data quality for analysis and automation and introduced Robotic Process Automation (RPA). Poor quality data can result in flawed decisions and outcomes. To ensure data used in automation is accurate, consistent, and up-to-date, businesses can use data quality management processes like data cleansing, validation, and profiling. RPA is an ideal technology to automate time-consuming and error-prone tasks like data entry, extraction, report generation, invoice processing, and customer service. This can improve efficiency, reduce errors, and free up employees to focus on higher-value tasks, ultimately leading to improved customer satisfaction and revenue.

In this class, we will begin understanding how data quality affects automating using a guessing game bot as an example. Even though we can construct a bot, if we don't anticipate certain errors, then our bots will fail, which will cause our automation solutions to fail. This case sets the stage for building robust bots as part of our automation solutions.

Materials and Suggested Seating:

Case: Our First UIPath Bot
Slides for this class have been archived for this quarter.
Automation Tools: UIPath
Suggested in-class seating: There is no recommended seating for this class. Working on the bot and any class discussions will be undertaken in pairs or small groups pairs..

Suggested Pre-Class Preparation:

1. If you didn't get a UIPath license in the prior class, please obtain this before class, see: UIPath Academic Alliance Licensing Page.
2. You are encouraged to read through the Our First Bot Case before class.

Class Plan:

1. First, we will have a very brief review giving some color to the Review and Extension section above. Expect pair discussion and PollEverywhere polls for this material.
2. We will begin working on the bot in UIPath. In most cases for desktop-level automation we will use a Flowchart tool as our starting point.
3. We will continue to work on our process using a combination of UIPath Activities including, Message Box Activity (for prompting the user), Assign Activity (to generate data), and Flow Decision Activity (a way to test logic in our workflow).