Automation Anywhere, Inc.

Ranjna Garg Goyal

Niraj Radhakrishnan Kunnumma

Narayanan Seshadri

Radek Hecl

Niveditha Nataraj

Improved techniques for producing digital fingerprints to represent steps in a recording of user-initiated events with respect to one or more application programs. A digital fingerprint can be formed for each step in a recording by encoding a plurality of attributes for each step in the recording. The encodings can be determined differently for different attributes of the steps. Also, the particular attributes to be used together to form a digital fingerprint can be chosen differently to adjust effectiveness in matching such step with other similar steps. In one embodiment, the digital fingerprint for a given step in a given situation can dynamically generated, such as a request to a server to generate a fingerprint.

Dynamic fingerprints for robotic process automation

Improved techniques for examining a plurality of distinct recordings pertaining to user interactions with one or more software applications, where each recording concerns performing at least one task. The examined recordings can be processed such that the recordings can be organized and/or rendered in a consolidated manner which facilitates user's understanding of higher-level operations being performed by the examined recordings to carry out the associated task. As one embodiment, a robotic process automation system can, for example, operate to: acquire a plurality of recordings, the recordings include a series of user interactions with one or more application programs, the recordings being acquired via the robotic process automation system; identify repeating sequences of user interactions within the recordings; select at least one of the identified repeating sequences that occurs more often; create a first level pattern for the selected at least one of the identified repeating sequences; and associate a descriptive label to the first level pattern created for the selected at least one of the identified repeating sequences.

Repeating pattern detection within usage recordings of robotic process automation to facilitate representation thereof

Improved techniques for combining a plurality of distinct recordings pertaining to user interactions with one or more software applications each of which concerns performing a task. Then, the combined recording can be used to form a software automation process for performing the task in an automated fashion. Advantageously, the improved techniques enable a robotic process automation (RPA) system to capture multiple recordings, combine (e.g., merge) the recordings, and form a software automation process to perform repetitive tasks in an automated manner, all with little or no assistance from a user.

Automated software robot creation for robotic process automation

Improved techniques for examining a plurality of distinct recordings pertaining to user interactions with one or more software applications, where each recording concerns performing at least one task. The examined recordings can be processed such that the recordings can be organized and/or rendered in a consolidated manner which facilitates user's understanding of higher-level operations being performed by the examined recordings to carry out the associated task. Advantageously, the improved techniques enable a robotic process automation (RPA) system to recognize and represent repetitive tasks within multiple recordings as multi-level (e.g., hierarchical) patterns of steps, sub-tasks, or some combination thereof. In doing so, a RPA system can identify and define such patterns within recordings and can also accommodate variants in such patterns. The resulting multi-level representation of the recordings allows users to better understand and visualize what tasks or sub-tasks are being carried out by portions of the recordings.

Growing community of inventors

Santa Clara, CA, United States of America

Ranjna Garg Goyal