Using AI Data Explorer for instant data insights

Pardeep K
ServiceNow Employee
2 hours ago

AI Data Explorer for IRM 

Query your risk data in plain language — no report builder required. 

 

This is one in a series of articles exploring capabilities on the ServiceNow AI Platform that you may not be aware of but could be useful in improving your risk program.  

 

What it is and the problem it solves 

AI Data Explorer is a conversational analytics capability built into ServiceNow Integrated Risk Management (IRM). It lets you type questions—such as 'Show me open high-risk findings from the last 90 days'—and get structured results from IRM tables without writing GlideRecord queries or building reports. 

 

AI Data Explorer gives risk practitioners easier access to the data they need without depending on a developer or report designer. Invaluable information piles up in the platform but retrieving them for audits, board packs, or Risk and Control Self-Assessment (RCSA) reviews requires either pre-built reports or someone with query skills. AI Data Explorer removes that bottleneck. 

 

The AI Data Explorer application is available from the ServiceNow® Store. AI Data Explorer skills can be found under Now Assist skills for Data and Analytics.  

(view in My Videos)

 

How it works 

 Stage 1 — Entry trigger 

Starting from any list, data visualization, or document on the ServiceNow AI Platform, AI Data Explorer surfaces AI-generated insights that highlight trends, anomalies, and key callouts. The three entry points are: a list view (any ServiceNow table), a workspace panel (e.g. Service Operations, IRM), or a Platform Analytics dashboard widget via the "Analyze trend + AI" button. All three enable the user to access the same exploration session without leaving their current context.  

 

Stage 2 — Natural language question + goal 

The user types a plain-language question and sets an exploration goal — a brief statement of what they're trying to understand (e.g. "Monitor control failures by region"). Instead of relying on pre-built dashboards or waiting for a dedicated analyst, a user can ask a question like "Show me the average resolution time for high-priority incidents in London last quarter."  

 

The AI immediately processes this query, generates the necessary filters, and presents the data in an intuitive, visual format. The goal statement anchors the AI's subsequent analysis and, in the future, drives action recommendations.  

 

Stage 3 — Analytics query generator (Now Assist skill, LLM: Azure OpenAI) 

This skill is the natural language-to-query translator that takes the user's question and converts it into a structured data query against the target table. The right filters, fields, date ranges, and groupings are applied without the user having to use any GlideRecord syntax. This makes AI Data Explorer accessible to non-technical business users. 

 

Stage 4 — Analytics exploration skill (Now Assist skill, LLM: Now LLM LTS Service) 

The exploration skill executes the generated query against live ServiceNow data. It queries the actual records, not cached exports or pre-built datasets. AI Data Explorer has access to data across the enterprise and even third-party systems using the Workflow Data Fabric (WDF). You can reach beyond native ServiceNow tables if WDF connectors are in place. Results are returned as structured data ready for the next stage.  

 

Stage 5 — Analytics insights generation (Now Assist skill, LLM: Azure OpenAI) 

At this stage, raw query results gain meaning. The skill generates three things simultaneously: a written AI narrative (explaining what the data shows in plain language), an auto-chart (selecting an appropriate visualization — bar, trend line, pie — based on the data), and anomaly/pattern flags that highlight non-obvious findings.  

 

The results show standard analysis vs. extended analysis (pattern discovery), and follow-up questions that can help build context into the conversation. 

 

The two analysis modes include: 

  • Standard analysis — Directly answers a question with a chart and narrative 
  • Extended analysis — Digs deeper, scanning the data for patterns the user didn't ask about. Powered by the optional analytics hidden insights generation skill, AI is asking "What else should I be aware of about this dataset that I didn’t ask about?" 

 

Stage 6 — Exploration document 

The output isn't a one-time answer. It's a persistent, editable exploration document. Users collect, edit, and organize insights directly within their exploration, and work with colleagues to refine analyses, align on insights, and accelerate data-driven decisions in a shared space.  

PardeepK_0-1778691052895.png

Exploration includes the ability to: 

  • Edit and organize — Users can rewrite AI-generated text, reorder insights, and add their own commentary. 
  • Collaborate — Teammates can join the same exploration in real time. 
  • Action recommendations — AI suggests next steps based on the goal and exploration content. 
  • Summarization — The optional Exploration summarization skill generates a concise summary so those new to the conversation can get up to speed quickly. 

How to configure it 

1.    Enable the Now Assist skill for IRM — Navigate to Now Assist Admin › Skill Configuration. Find the IRM Data Explorer skill card and toggle it to Active. Confirm the status shows Running. 

 

2.   Assign roles — Assign two roles to every user who needs access: now_assist.user (grants access to the Now Assist panel) and sn_grc.data_explorer (scopes LLM queries to IRM tables only). 

 

3.   Validate LLM connectivity — Open “connection &[PK1]  credential aliases” (sys_connection_alias.list). Find the Now Assist LLM alias and click Test Connection. You need a 200 OK response. If it fails, your network team will need to whitelist the ServiceNow AI endpoint. 

 

4.   Verify IRM table scope — Confirm the IRM tables listed above are accessible in your instance scope. These are pre-configured; no manual table mapping is needed. 

 

Requirements 

Requirement 

Detail 

Where to confirm 

Platform release 

Xanadu or later 

System Properties › glide.buildtag 

IRM plug-in 

sn_grc — IRM Advanced entitlement required 

Plugins › sn_grc 

Now Assist SKU 

Now Assist for IRM, activated on the instance 

Now Assist › Skill Configuration 

Roles 

sn_grc.manager or sn_grc.user + now_assist.user + sn_grc.data_explorer 

User Administration › Roles 

LLM connectivity 

Instance must reach the ServiceNow AI endpoint (no proxy block) 

Connection & Credential Aliases 

  

ServiceNow resources 

0 Helpfuls
  • 69 Views