AISense overview
- Updated: 2020/04/17
AISense overview
AISense provides advanced capabilities and learning algorithms to make automation in all environments faster and more accurate.
A bot can only perform mouse clicks and keystrokes to automate tasks in remote desktop environments. This restricts the bot to using only image, coordinate, and OCR commands. The following are some of the issues that affect the accuracy of an automation task in remote environments:
- Automation platform receives an image of the application and does not have access to the UI elements.
- Screen resolution and scaling might not be the same on the host and client machines.
- The position of a UI element might change for web applications on the screen of client machines.
AISense provides AI techniques such as intelligent image recognition using computer vision to identify the UI elements of an application image. It automatically creates all UI objects from that image. It enables you to automate tasks on remote desktop environments. These objects are available for automating a task through MetaBot Logic, helping you save a considerable amount of time.
AISense enables you to automate tasks for:
- Applications that are available in a Citrix environment (Citrix XenDesktop and Citrix Virtual Apps).
- Applications that are accessed over Remote Desktop Protocol (RDP).
- A legacy application such as Delphi.
AISense uses computer vision to intelligently create dynamic linking between objects by determining their composition to deliver change-resilient automation. AISense enables you to accurately automate applications even when labels and text change their position.
You can use the Export Dataset command in MetaBot to extract a huge amount of data available in all the UI elements of an application. Similarly, the Import DataSet command enables you to insert a huge amount of data in all the UI elements of an application. You can use these commands for standard applications and remote applications. AISense enables you to extract or insert data in all the objects visible on the screen in a single scan of the application and reduce the execution time considerably.
- Auto linked: Objects, such as text boxes, are linked to another object during screen creation.
- Manual linked: The objects are manually linked by you.
- When using the Remote Desktop application to capture a screen, enable the full-screen option.
- When using the OCR command, ensure the DPI is set to standard or 100.
- Ensure that the occurrences are ordered from top-to-bottom and left-to-right.