Predictive intelligence

[poonamv]

Hi Team,

I am looking for some use cases of predictive intelligence for ITSM,if anyone has Implemented,could you please share more details.

Thanks in Advance

[Dr Atul G- LNG]

Hi @poonamv 

• Show suggestions for relevant articles.
• Assign, categorize, and prioritize tasks.
• Detect major incidents.
• Recommend case resolutions.
• Prevent duplicate articles and ideas.
• Detect phishing attempts.

https://www.servicenow.com/community/developer-articles/predictive-intelligence-made-simple-a-use-ca...

[Vutikonda Vinay]

Hi,

Someone already got helpful earlier post you can refer the same.

https://www.servicenow.com/community/itsm-forum/looking-for-real-time-benefits-and-use-cases-of-pred...

[Ehab Pilloor]

Hi @poonamv,

Type: Classification

Usecases:

Predict Assignment group field from the Short description.
Predict the Category field from the Short description.
Predict the Configuration item field from the Short description.

Type: Similarity

Usecases:

Recommend similar major incidents which the current incident can be linked to.
Recommend similar incidents based on the Short description and Description fields, to help with incident investigation process.
Recommend similar closed incidents based on the Short description and Description fields, to help with incident investigation process.
Recommend similar resolved incidents to help with incident investigation process.

Regards,

Ehab Pilloor

[Aniket Chavan]

Hello @poonamv ,

I’ve implemented Predictive Intelligence in ITSM primarily using the classification framework, and it’s been really effective for use cases like auto-assigning the assignment group or predicting the category based on the short description. I’ve trained and deployed multiple models end-to-end in this area.

If you’re just getting started, I’d suggest beginning with the OOB assignment group prediction for incidents. It’s easy to set up and gives pretty good results right away. Once you’re comfortable, you can fine-tune the predictions — even adjust the confidence threshold per group if needed.

Also, one suggestion — I’d recommend starting with the out-of-the-box solutions first. They’re solid, easy to implement, and give you a good understanding of how PI works behind the scenes. Once you’re confident with that, you can start building custom models tailored to your or your client’s specific needs.

Also, one thing that really helped us was reporting on the PI prediction data itself. The system logs what value PI predicted and whether the end user changed that value. So, by analyzing that table, we could understand how accurate the model was in real usage and use those insights to further improve our solutions. It’s super helpful in tuning the model to match how your users actually work.

Just for context, PI supports three main frameworks:

• Classification – to set field values like category or assignment group

• Similarity – to suggest resolutions based on similar past incidents

• Clustering – to detect patterns or group similar records (like for major incident detection)

🔹 Please mark ✅ Correct if this solves your query, and 👍 Helpful if you found the response valuable.

Best regards,
Aniket Chavan
🏆 ServiceNow MVP 2025 | 🌟 ServiceNow Rising Star 2024