top of page
  • Writer's pictureAlex Carter

How can Deep Reasoning be applied on Process Mining?

Updated: Aug 10, 2022


deep reasoning on process mining

Process mining generates multiple data sets which need analysis to determine the root cause of the process deviations. Deep reasoning could be a way to automate this process of identifying the root cause analysis. There could be multiple factors resulting in a process deviation. The reason why and when this factors come up could be conditional, behavioral or situational etc. The idea of deep reasoning is to predict these events which may occur in the future by applying AI\ ML techniques.

Deep Reasoning builds up its own knowledge graph based an AI\ Machine learning based technique used to learn the root cause for process deviations automatically from results of process mining. A cognitive model will constantly learn deviations from various process cycles and analyze the root cause for it. After sufficient learning, it would be in a position to automatically predict the process deviations, which may occur in the future based on its past knowledge gathered from previous process cycles.

Deep learning technique are employed to constantly learn and build a knowledge graph of the possible root cause for the deviations. Based on the gathered knowledge the system can try to predict if a particular process will have deviations and if so what would be the possible root cause.


Based on the accuracy of the prediction the system can hint the operator about a possible deviation and give justification for this. The operator can take a leverage on this prediction and already deploy suitable measures (RPA bots) so the deviation would not even take place. This is pre-emptive approach and has a lot of potential to reduce unwanted costs.

This prediction could be used to generate alerts and counter measures could be taken so we could avoid the upcoming process deviation as much as possible. An automation system or RPA could be setup to keep track of this prediction events and take counter actions to mitigate any process deviations as much as possible. This system would constantly monitor your process flow and take corrective actions when needed 24\7 without any manual interventions.

Deep reasoning not only generate predictions but also give a reasoning for its prediction. This is different from other deep learning techniques were its very hard to understand the justification for a prediction. With deep reasoning, an operator can look at the root cause justification for any prediction and judge if the counter measures are valid in this case.

Based on the root cause justifications the operator can setup correct RPA, which can take corrective actions to avoid process deviations. Different root causes could mean different actions to be taken. Multiple RPAs could be configured based on the different corrective actions needed. A rule-based approach could be setup which would trigger different RPAs based on the root causes registered. With this approach even complex process flows can be monitored and simple rule based action list can be created which is easy to understand and setup by the operator.

Deep Reasoning has a long way to go but it’s the future of process mining as we see.


Contact us today if you want to increase productivity and improve optimization with process mining along with identifying opportunities for automation.


Comments


bottom of page