Rozmin Parpia
ServiceNow Employee
ServiceNow Employee

 

RozminParpia_0-1745948294551.png

ServiceNow Process Mining allows you to filter the mined data and process map. You can convert a filter into finding (pattern) quickly and easily. This helps you:

 

  • Add the finding to multiple projects by simply clicking on ‘add from library’
  • No need to remember the filter criteria
  • Save time spent on recreating same filters on different projects
  • One-time setup making it easier for teams to detect the pattern on their mined projects
  • Easy way to compare the findings across projects

 

Filters, Findings, and its types

 

Before we dive into the steps required for converting a filter into finding, let’s understand these concepts.

 

There are 2 types of filters:

 

RozminParpia_1-1745948294552.png

 

 

Conditions Filter

 

It uses fields from the table mined or its related tables as filter criteria. In the example below, we have filtered for ‘active’ incidents.

 

RozminParpia_2-1745948294554.png

 

Transitions Filter

 

It uses transitions on the process map as filter criteria. It is applied on the 'analyst workbench' to refresh the process map with specified criteria.

 

RozminParpia_0-1745951914861.png

 

In the example below, we have filtered for incidents that have ‘state’ as ‘in progress’, directly followed by ‘state’ as ‘awaiting caller info’. 

 

RozminParpia_3-1745948294560.png

 

You could further add ‘constraints’ to specify how long the records took for the transition. In the example above, minimum duration is 2 days. The filter will then pick incidents that took 2 days or longer to go from ‘in progress’ to ‘awaiting caller info’ state.

 

There are 2 types of findings (patterns):

 

Rule based finding

 

It exposes records based on rules that you set up e.g. show me records where incident state is ‘resolved’ for more than 7 days, eventually followed by ‘closed’. If you notice, the rules in this finding (pattern) seem very much like the 'transition' filter criteria described above.

 

Automated finding

 

It uses preset patterns provided in the plugin to expose the bad process behaviours. All you need to do is specify the ‘field’ you would like to detect that pattern on e.g. state, assignment group.

 

Examples of automated findings are:

 

  • Ping pong: records going back and forth between 2 activities e.g. from group A to B and back to A.
  • Rework: an activity in the process needs to be redone
  • Extra step: deviation in common route by one extra step e.g. ‘new’ to ‘in progress’ to ‘resolved’ versus ‘new’ to ‘assigned’ to ‘resolved’

 

How to convert a filter into finding

 

Now that you know the basics of what filters, findings and its types are, you can easily convert a ‘transition’ filter into a ‘rule based’ finding.

 

We recommend performing this action on the ‘Process Configuration’ instead of your project so that you reap full benefits of leveraging the finding (pattern) across multiple projects. If you create a finding on your project instead you will not be able to apply it to other projects.

 

We will now learn to setup a rule-based finding: Longer to route than resolve. It helps expose incidents / records that took 2 days or longer to route to the correct group but took only 30 mins to resolve. This help us improve the processes by automating the steps, coaching our teams, creating knowledge articles or using Virtual Agent topics.

 

Steps to configure the finding

 

Go to the ‘process configuration’ and open the table you’re interested in creating the finding (pattern) on as shown in the visual below.

 

RozminParpia_4-1745948294565.png

 

 

           

RozminParpia_5-1745948294566.png

 

From ‘finding definitions’ tab click on ‘New’. It is recommended to ‘use finding builder’ if you are in the recent release to ensure you are assisted through a guided setup to create findings.

 

RozminParpia_6-1745948294568.png

Message is the name of the finding i.e. longer to route than resolve. Category is ‘performance’ as this finding affects the performance of the process. Then proceed to ‘configure’.

 

RozminParpia_7-1745948294574.png

 

Set the criteria here that you would have used in the ‘transition’ filter on your project. State is ‘in progress’ eventually followed by state is ‘resolved’. Occurrence is ‘All’ for ‘in progress’ whereas it is ‘Last’ occurrence for ‘resolved’ state.

 

Then add constraints and specify the max duration as 30 mins. After adding constraint, go to ‘Add chain’. Fill out the criteria for the second chain as shown below.

RozminParpia_8-1745948294581.png

 

You have now setup the rule based finding - ‘longer to route than resolve’.

 

Next step is to go to your project from the workspace, edit it and on step 3: improvement opportunities, on rule based finding, ‘add from library’ (visual below).

 

RozminParpia_9-1745948294584.png

 

You will see that the new finding exists in the library as shown below. Apply it to your project and mine the project.

RozminParpia_0-1745949578116.png

 

After mining, on the ‘Insights and Summary’ page you will see your finding (pattern): longer to route than resolve as depicted below.

 

 

RozminParpia_12-1745948851511.png

 

Guidance

 

Please bear in mind:

 

  • The finding (pattern) will not show up on the summary page if your data does not have that pattern i.e. your process does not suffer from this anomaly.
  • Data fields in your instance might have been relabelled or custom fields created. So, please ensure you are picking the correct fields for the pattern setup after consulting with your system administrator.

*Full video recording and demo from the academy session is here.