Process Simulation

Challenge

Optimizing a process is not an easy task. Too many variables, inputs, and steps involved affect each other. Uncertainty in changes can lead to an unintentional decrease in overall process performance.

Do you ask questions like:

• What would happen if I had a bigger team in the Accounting department?
• How much cost would I save with 3 RPA bots taking over the Invoice Retyping activity?
• Would my customer journey be faster if I rerouted the returning orders to pass our overloaded registry office?

Solution

Immediately see how changing the variables translates into the process performance, map, and statistics, and get the answers to your what-if questions.

Simply apply Minit Process Simulation on an existing process and custom-set variables affecting it, such as resources dedicated to a task, manhour rate, working hours and days, distribution of process starts, task duration, and others.

To make the simulations more accurate and precise, we have added the option to use machine learning models in each activity. 

How Machine Learning (ML) Works in Simulations

Without ML in Simulations, each activity decides on random what the next path will be for any given case. This can cause irregularities in the newly generated log in many cases.

Imagine you have a process that consists of two variants. One is ABCE and the other is ACDE. So if a simulation is simulating such a process, variants such as ACE or ABCDE will occur, but this is not what the original process was supposed to look like.

So, if you include case attributes in simulation, you can teach the model that cases with certain values of case attributes are taking the first variant and cases with different case attributes are taking the other variant. This will result in precise simulations that will return generated logs matching reality.

Minit also allow using ML for choosing resources in every activity based on the original log and the duration of a case in any given activity.

Control

Once the simulation confirms that the assumed changes deliver the desired results, you know what changes need to be applied and how to optimize the process. This allows you to make a risk-free decision, based on data, not assumptions.