AI in Robotiq.ai
AI is contained in the name of our product so how exactly it is used and how does it work? On one side we use AI to simplify how processes are developed and on the other side to provide more robustness to the robots.
AI is quite a generic term, but in our case, we use machine learning and shallow neural networks to provide cognitive capability to the robot. We call this component the Vision Center (VC) and it is used by the robot as well as the recorder application. Role of VC is to classify application elements and locate them on the screen. What does it all mean?
It means that our robots recognize the elements of the application (buttons, textboxes, dropdown menus, etc.) visually instead of using so called selectors. Selectors require technical knowledge and are not available on all application platforms such as Java or Oracle Forms. Also, sometimes they can be unreliable and cause robots to crash.
How does Vision Center work in practice? For example, when using Click step instead of having to work with selectors, we can just upload an image of the button. User can easily snip the image, or the recorder generates it. If the look and feel of the button changes the user can easily replace this image with a new one. It’s very intuitive and the user immediately knows what the robot will perform in this step. This is the simplicity part.
Robustness feature comes into play when the robot executes processes. When robot comes to a certain step and needs to click on a control it takes a screenshot of the application and the image of the control from the process and hands it over to VC. VC searches for the control and returns the coordinates to the robot which then performs the click. Using different techniques, we can locate the control even if the screen resolution is different, control is located on a different place or even if it changes look and feel. This is all achieved without the need to modify your process. And that is the power of AI in our platform.
But this is just a first use of AI in our platform. In the following releases we will introduce various features such predictive process design or scheduling assistant the further simplify how we develop and maintain the robots!