Method for controlling a process using a program product
View Patent ↗Methods are known for controlling processes using control software which keeps different control elements ready at the same time. In this respect, the process of suppressing or putting control elements out of service if the control elements are not available is likewise known. The invention is to be improved, however, in that control elements and decisions are given to the user such that an intuitive use is facilitated, and the user is guided through the control of a process. For this purpose, a process control is carried out via a program product, which is represented as a chat. The user and the program product act as communication participants, wherein the planned process is controlled by control commands which are requested in the course of the chat and entered or selected in response thereto.
1 . A method comprising:
generating a first chat session responsive to a control signal that invokes a workflow in a workflow database the control signal generated responsive to a fault in an enterprise process, the first chat session having at least a first real participant and at least a first virtual participant controlled by the workflow, the first virtual participant uniquely associated with the enterprise process;
uniquely assigning a first chat bot, by a processor over a distributed network, to the first real participant, the first chat bot using semantic context and any prior predictions to predict a chat interaction by the first real participant and comparing the predicted interaction with an actual interaction and refining future predictions based on an accuracy of the predicted interaction;
uniquely assigning a first parser to the first chat bot, the first parser, parsing messages in the first chat session including messages submitted by the first real participant;
modifying a first knowledge base, uniquely assigned to the first chat bot, over the distributed network within a storage resource coupled to the distributed network based on the accuracy of the prediction;
modifying the first knowledge base over the distributed network to include a linguistic style of the first real participant based on the parsing of the messages submitted by the first chat participant;
repeating the prediction, refining, and modifying for a plurality of interactions in the first chat session; and
timestamping each interaction of the first chat session.
2 . The method of claim 1 further comprising:
generating a second chat session responsive to a subsequent control signal that invokes the workflow in the workflow database;
substituting the first chat bot in place of the first real participant to interact with the first virtual participant during the second chat session without input from the first real participant.
3 . The method of claim 1 further comprising:
evaluating response times for participants in the workflow by analyzing a plurality of time stamps from the first chat session to determine a throughput value for the workflow.
4 . The method of claim 1 further comprising:
calculating from a plurality of time stamps an average response time of the first real participant.
5 . The method of claim 1 wherein a second real chat participant participates in the first chat session further comprising:
uniquely assigning a second chat bot to the second real chat participant, the second chat bot using semantic context and any prior predictions to predict a chat interaction by the second real chat participant and comparing the predicted interaction with an actual interaction and refining future predictions based on an accuracy of the predicted interaction.
6 . The method of claim 5 wherein predicting the chat interaction comprises:
parsing messages within the first chat session to generate ordered message data; and
comparing, in the second chat bot, the ordered message data with prior learned responses to infer a probable response based on similarities between the ordered message data and ordered message data from at least one prior chat session.
7 . The method of claim 1 further comprising:
assigning a first parser to the first virtual participant; and
parsing the messages in the first chat session to directly generate control commands within the workflow.
8 . The method of claim 7 further comprising:
issuing the control command from the first virtual participant through a process interface that provides access to at least one process external to the first chat session.
9 . The method of claim 5 wherein each of the first and the second chat bots are assigned a unique knowledge base at creation, the method further comprising:
independently building each unique knowledge base from the predictions and the prediction accuracy of the assigned chat bot.
10 . The method of claim 9 further comprising:
collectively monitoring the first and second chat bots to compare knowledge bases and relative prediction accuracy.
11 . The method of claim 1 further comprising:
assigning a first parser to the first virtual participant; and
parsing the messages in the chat session to trigger a further action or command within the workflow.