US 12,456,083 B2
Systems and methods for automation discovery recalculation using dynamic time window optimization
Eran Roseberg, Hogla (IL); Yaron Moshe Bialy, Madrid (ES); and Yuval Shachaf, Even Yehuda (IL)
Assigned to NICE LTD., Ra'Anana (IL)
Filed by NICE LTD., Ra'anana (IL)
Filed on May 12, 2022, as Appl. No. 17/742,944.
Prior Publication US 2023/0368104 A1, Nov. 16, 2023
Int. Cl. G06Q 10/063 (2023.01); G06Q 10/0631 (2023.01); G06Q 10/0633 (2023.01); G06Q 10/0637 (2023.01); G06Q 10/10 (2023.01)
CPC G06Q 10/0633 (2013.01) [G06Q 10/06312 (2013.01); G06Q 10/0637 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for identifying ordered sequences of processes comprising routines, each routine comprising a series of actions input to a computer, the method comprising using one or more computers:
gathering, by a data collection software executed by a computer processor, a plurality of low-level user action data items describing input by a user to one of a plurality of software programs via a desktop system, each low-level user action data item comprising an input type description, and screen window information;
determining, using sequential pattern mining, a plurality of routines, wherein each of the routines includes a series of user actions, the series of actions recurring in the low-level user action data items;
for a given process comprising an ordered sequence of a plurality of the routines:
recalculating, for a plurality of instances of the given process in the low-level user action data items, one or more transition probabilities for one or more pairs of routines, each pair of routines connected by a transition, wherein each instance of the plurality of instances was used for discovering the given process in the low-level user action data items;
updating the given process based on the recalculated transition probabilities; and
automatically executing, by a bot, the updated process, wherein the executing is performed at runtime, and wherein the updated process is compiled into a code.