Clarification on AIOps LEAP Incident Clustering: Role of ML, Skills, and LLM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
3 weeks ago
Hi Team,
I have a question regarding how AIOps LEAP clustering works in the backend.
According to the ServiceNow documentation, AIOps LEAP performs incident clustering using machine learning (ML), skills, and LLM models. However, I am trying to understand where exactly these components are implemented in the backend and how they contribute to the clustering process.
While exploring the instance, I found a record under Predictive Intelligence, and I am trying to confirm whether this record is actually used for AIOps LEAP incident clustering. I have attached SS2 as a reference for the ML record I found.
From my current understanding:
Machine Learning clusters incidents based on fields such as short description, work notes, and resolution notes.
Skills might be used for topic generation within the clusters.
However, I was unable to find clear documentation confirming this architecture, so I would appreciate any clarification.
For reference:
SS1 shows the clustering flow mentioned in the ServiceNow documentation.
SS2 shows the ML record I found under Predictive Intelligence.
SS3 shows the skill that appears to generate topics for the clusters.
Could someone please confirm:
Whether the Predictive Intelligence record I found is used for AIOps LEAP incident clustering?
What exactly is the role of ML, skills, and the LLM model in the clustering process?
If there is any official documentation or articles explaining this architecture, it would be extremely helpful if you could share them.
Thank you very much for your time and support.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
2 weeks ago
Here is my small Input .
ServiceNow AIOps LEAP uses ServiceNow® large language model (LLM) plus machine learning clustering on closed IT incidents to identify automation opportunities.
It “leverages AI and machine learning clustering capabilities” and automatically groups similar incidents based on patterns and signatures (auto‑categorization using Now LLMs).
Those incident clusters create a knowledge foundation that “continuously improves over time,” and LEAP uses GenAI (LLM) to extract resolution steps from historical resolutions tied to the clustered issues.
LEAP then converts the generated resolution steps into actionable playbooks—including the documented flow where you can Generate resolution steps and create a playbook (Workflow Studio / Now Assist LLM directions prefilled).
Finally, LEAP supports prioritization (based on estimated savings/impact) and provides a central dashboard to track automation performance/ROI (cost savings & effectiveness of playbooks).
So This whole Feature is a Mix of PI and Gen AI where Both Are Complementing each Other.
Below Link will also give you some Pointed Answers.
Regards
RP
