AI-Powered Incident Summarization and Skillset Extraction

[siddheshagn]

Step 1: Define the Use Case

• Goal: Extract and summarize key topics from incident records.
• Output: Store summaries and keywords in a custom table for reporting.

---

Step 2: Create a Custom Table

• Navigate to System Definition > Tables
• Create a table named incident_keywords
  • Fields: Incident Sys_ID, Keyword, Summary

---

Step 3: Build an AI Agent

• Go to AI Agent Studio > Create and Manage
• Define:
  • Name: Incident Summarizer Agent
  • Role: “Extracts key topics and summaries from incident descriptions”
  • Instructions:
    1. Retrieve incident details
    2. Analyze text using AI
    3. Match or generate keywords
    4. Insert results into incident_keywords table

---

Step 4: Add Tools to the Agent

• Script Tool – Fetch Incident: Gets Short Description and Description
• Script Tool – Get Existing Keywords: Queries known keywords
• Record Operations Tool – Insert Keywords: Adds new records to the custom table

---

Step 5: Configure Trigger

• Use a Record-based trigger:
  • Trigger when a new incident is created or updated
  • Agent runs in the background

---

Step 6: Activate and Test

• Click Activate in AI Agent Studio
• Use the Test screen:
  • Provide an incident number
  • Observe how the agent:
    • Pulls description
    • Extracts keywords
    • Inserts them into the custom table

---

Optional Enhancements

• Use Google Gemini or OpenAI via Now Assist for deeper summarization 
• Auto-update Work notes with summaries
• Recommend Knowledge Articles based on extracted topics