I recently had to help set up Predictive Intelligence for PPM for a client. I realized it wasn't just turning on a plugin, and it's good to go. There are some configurations needed to get it to work. I wanted to make life easier for the rest of the community with this guide and highlight a common issue.
If you don't know what Predictive Intelligence for PPM is all about, in its simplest form, it helps reduce the creation of duplicated entries for Project, Demand, and Ideas. It does this by looking for "keyword(s)" in different fields within existing Projects, Demands, and Ideas. It provides the end-user with a list of "related" items for the record they are trying to create. The end-user can now compare the record they are trying to create with the list of existing ones.
Another feature it can provide is the ability to find other Projects, Demands, and Ideas that can have similarities. The end-user can then look at those Projects, Demands, and Ideas and use the old information from the records on the new record.
Predictive Intelligence for PPM only comes with the ITBM Professional package. If you want to know more about pricing, please reach out to your ServiceNow Account team.
Click on the links to learn more about Predictive Intelligence and Contextual search features. Here is the FAQ for Predictive Intelligence as well.
Before we start, you will need a user with the Admin role.
Step by step guide to set up the Predictive Intelligence for PPM ( Demand, Project, and Idea).
Step 1 - Turn on the Predictive Intelligence plugin -
PPM - Predictive Intelligence for PPM - com.snc.ppm_ml
Idea - Predictive Intelligence for Ideation - com.snc.innovation_management_ml
*If you do not have the "Predictive Intelligence Contextual Search" plugin turned on, it will be turned on when the above plugins are activated. It is required for them to work.
Step 2 - Once the plugin(s) has been loaded, you will have to train the ML Solutions. This step is the same for Project, Demand, and Idea solutions. You have to go to the Solution Definitions to create the ML solutions. Go to Predictive Intelligence > Similarity > Solution Definitions. These definitions are OOTB that SN provides. You can create your own. Here is the documentation for setting up a new one.
For this example, I will use the Similar Projects record (this step is the same for the Demand and Idea records.) Once you open the record, you will have to click on the "Update & Retrain" button. This creates an ML Solution and is needed for Predictive Intelligence to work. Without this step, the PI will not provide the user with a list of related records.
Once the "Update & Retrain" button is clicked, it should create a new ML Solutions record; you can see it in the related list tab. Keep an eye on the State and Progress field. These two fields are a good indicator of possible issues they may come up with the training. Getting the ML Solution back from the SN server will take time, don't be alarmed if you don't get instant results. The time it takes to get a result varies from user to user. This example ML Solution took about 3 mins to get back from the SN server.
Here is the configuration tips documentation from SN. It can help with troubleshooting if any issues come up. You can also look in the SN Community to help with issues.
*The one issue that came up for me was that the ML Solution was stuck in the "waiting for training" for over a day. If that happens to you, please check the "glide.servlet.uri" in your instance's system property is set up with the correct instance URL. This is usually seen if you clone down your instances from Prod. I set the "glide.servlet.uri" property with the correct URL, resolving the issue.
Once the ML Solution comes back, PI should be working on Project, Demand, or Idea. Here's how to tell, Create a new Project, Demand, or an Idea. You should now see a list of "similar" records to the one you are trying to create.
Demand
Idea
I hope you find this useful! If you think I've missed something, please let me know so I can get it added.
1 Comment
