question Re: Enhance Discovery With Artifical Intelligence? in ITOM forum

Enhance Discovery With Artifical Intelligence?

Tom Rausch — Fri, 16 Jan 2026 01:07:01 GMT

What are some ways I can enhance ServiceNow Discovery with features provided by artifical intelligence?

Re: Enhance Discovery With Artifical Intelligence?

Matthew_13 — Fri, 16 Jan 2026 05:05:39 GMT

Hi Buddy,

Some ways I see great use of it are:

Improve CI quality by normalizing names, correcting CI classes, and flagging duplicates
Smarter reconciliation by identifying likely duplicate or conflicting CIs
Detect gaps and failures in Discovery runs (missing data, failing patterns, bad credentials)
Change and drift detection by flagging unusual CI or relationship changes
Assist service mapping by suggesting relationships and ownership where data is incomplete

Simply put, AI helps clean, validate, and interpret Discovery data, reducing manual CMDB work and improving accuracy.

@Tom Rausch - Please mark Accepted Solution and Thumbs Up if you found Helpful 🙂

Re: Enhance Discovery With Artifical Intelligence?

Tom Rausch — Thu, 22 Jan 2026 01:24:53 GMT

Hello, @Matthew_13

I agree. AI can provde benefits such as improver CI quality and smarter reconciliation.

Here is my question to you: What tools in ServiceNow can provide these benefits and improvements?

Re: Enhance Discovery With Artifical Intelligence?

Matthew_13 — Thu, 22 Jan 2026 01:57:02 GMT

Good question. In ServiceNow, these benefits don’t come from one single “AI tool” but from several capabilities working together.

CMDB Health and Data Manager help improve CI quality by highlighting missing data, incorrect classes, and relationship issues. The Identification and Reconciliation Engine (IRE) handles smarter reconciliation by preventing duplicates and resolving conflicts across data sources. Discovery and the Discovery Health Dashboard surface gaps like failed scans, bad credentials or incomplete patterns. AIOps and CI anomaly detection help flag unusual CI or relationship changes that can indicate drift or data issues. Service Mapping supported by ML helps infer relationships and ownership when information is incomplete.

So basically AI isn’t replacing CMDB discipline—it’s augmenting Discovery, reconciliation, and governance to reduce manual effort and improve trust in the data.

@Tom Rausch - Please mark Accepted Solution and Thumbs Up if you found Helpful 🙂

Re: Enhance Discovery With Artifical Intelligence?

Tom Rausch — Tue, 10 Feb 2026 01:21:05 GMT

GitHub Copilot can write ServiceNow Discovery Patterns. The prompt is natural language.

Re: Enhance Discovery With Artifical Intelligence?

Tom Rausch — Tue, 10 Feb 2026 01:23:10 GMT

GitHub Copilot indicates it can analyze ServiceNow Discovery Logs. I have not tested this capability, but apparently the capability is present.

Re: Enhance Discovery With Artifical Intelligence?

Tom Rausch — Tue, 10 Feb 2026 22:33:15 GMT

I asked GitHub Copilot, "How can GitHub Copilot help me automte operations for ServiceNow Discovery?"

Here is the response from GitHub Copilot.

GitHub Copilot

GitHub Copilot can help automate operations for ServiceNow Discovery in several ways, primarily by assisting you in writing scripts, integrations, or workflows that interact with ServiceNow’s Discovery APIs and tools.

Here’s how Copilot can be beneficial.

Script Automation

ServiceNow Discovery often requires writing custom scripts (in PowerShell, Python, or JavaScript) to automate tasks such as triggering discoveries, collecting configuration items (CIs), processing discovery results, or handling errors.

Copilot can assist by:

Suggesting code snippets for ServiceNow REST API calls.
Completing Python or PowerShell scripts to automate scheduled discoveries.
Providing boilerplate code for processing and parsing discovery results.

Workflow Integration

If you need to integrate ServiceNow Discovery with DevOps tools (like GitHub Actions or Jenkins):

Copilot can generate workflow files for orchestrating ServiceNow discovery via REST or MID Servers.
Copilot can create scripts to fetch discovery statuses, schedule new scans, or update CMDB from discovery results.

Error Handling & Logging

Automating error handling and logging is crucial for reliable operations:

Copilot can help write robust try-catch logic to handle various discovery API errors and log them in a suitable format.
It can suggest logging strategies compatible with ServiceNow and external monitoring tools.

API Integration

Most ServiceNow automation requires interacting with ServiceNow REST APIs (e.g., ServiceNow Discovery API documentation)

Copilot can help generate code to authenticate, query, and manipulate discovery jobs and results.
Copilot can prompt best practices for pagination, token management, and result parsing.

Custom Discovery Patterns

If you’re developing custom Discovery Patterns:

Copilot can assist with creating XML or JSON definitions and mapping logic, streamlining configuration and troubleshooting.

Infrastructure as Code

For organizations using IaC:

Copilot can help automate CMDB updates based on infrastructure changes, ensuring they are discovered and recorded in ServiceNow.

Example

Suppose you want to automate discovery job creation via ServiceNow’s REST API in Python:

import requests url = "https://yourinstance.service-now.com/api/now/table/discovery_job" headers = { "Content-Type": "application/json", "Authorization": "Bearer YOUR_TOKEN" } payload = { "name": "Weekly Network Discovery", "type": "network", "schedule": "weekly" } response = requests.post(url, json=payload, headers=headers) print(response.json())

create_discovery_job.py

Copilot can autocomplete, suggest improvements, and even generate documentation or README files for these operations.

Summary

GitHub Copilot helps automate ServiceNow Discovery operations by expediting script development, integration with external tools, error handling, API usage, and pattern creation—making IT operations more efficient and less error-prone.

Re: Enhance Discovery With Artifical Intelligence?

Christopher Hub — Wed, 11 Feb 2026 22:30:17 GMT

I have done pretty heavy testing of GPT, CP, Claude, and Grok for parsing and triage of automation_error_message table, and it's a rich mess of hallucination despite quite a bit of training effort. I similarly find that AI hallucinates heavily in the design of discovery solutions because it can't span knowledge of platform, knowledge of target devices, and discovery strategy. I'd love to hear that someone else has found traction, but so far I'm not seeing it. You can get very specific help with tactical issues from AI, but it falls on its face with complex work.