Quick Overview
AI Agent Advisor analyzes your operational data to identify automation opportunities and match them to ServiceNow AI Agents. This guide covers how it works, when to use it, how to interpret results, and how to get started.
1. Introduction to AI Agent Advisor
What is AI Agent Advisor?
AI Agent Advisor is a ServiceNow AI solution that analyzes your operational data to identify and prioritize automation opportunities, then maps those opportunities to the most suitable AI Agents, workflows, and tools available on the Now Platform. It transforms your incident and case data into a clear, data-driven automation blueprint.
It is currently available as an Innovation Lab feature, with General Availability planned for May 2026.
What does it do?
AI Agent Advisor operates across three sequential phases:
Mine
Uses AI-based clustering to group similar incidents or cases, extract intents, and identify resolution steps that are consistent across the group.
Match
Prioritizes opportunities by volume of related record and maps each resolution step to existing OOTB and custom ServiceNow AI Agents, Workflows, and Tools, generating an actionable solution blueprint with confidence scores.
Make
Provides native integration with AI Agent Studio to clone, create, and deploy AI Agents.
How does it work?
AI Agent Advisor is powered by the Group and Action Framework (GAF), ServiceNow's platform-level AI clustering engine.
There are four stages in the process:
|
1
|
Intent Extraction Each incident or case record is converted to a representation that captures the core issue. |
|
2
|
Embedding + Dimensionality Reduction Records are vectorized, and a process lessens the high-dimensional vector space, enabling efficient clustering. |
|
3
|
Clustering Density-based clustering groups semantically similar records into labeled automation opportunity categories. |
|
4
|
Agent/Tool Matching Each resolution step is matched against a catalog of OOTB ServiceNow Agents and Tools
|
2. When Should I Use AI Agent Advisor?
Requirements:
AI Agent Advisor produces the most valuable output when the following conditions are met:
| Requirement | Guidance |
|---|---|
| Record Type | Incidents and CSM Cases at present. |
| Volume | Minimum 500 resolved or closed records. More data produces better results. |
| Record Age | More recent records will yield opportunities relevant to your current state. The default filter retrieves records that have been resolved or closed within the last 365 days, but this is configurable. |
| Data Quality | Records should have populated work notes and/or comments. Records lacking detail increase the outlier rate, thus reducing the effectiveness. |
| Release | Zurich/Yokohama+ |
| License | Active Now Assist license |
AI Maturity Stages
One can derive value from AI Agent Advisor at any stage of their AI maturity - see some example scenarios below:
Getting started
You have a license that includes AI Agents, but have not yet identified which AI Agents to prioritize. AI Agent Advisor removes the cold-start problem by surfacing the biggest impact opportunities directly from your data.
Expanding coverage
You have deployed a few OOTB agents and want to identify where to go next. AI Agent Advisor finds the next-best opportunities from your actual operational data.
Demonstrating ROI
You want data to justify additional agent deployments to stakeholders. AI Agent Advisor provides volume-based prioritization tied to your real incident patterns.
Use Cases
Initial Deployment
Run AI Agent Advisor after you first install a Now Assist plugin to drive the first wave of agent deployments with clear business justification.
- Receive a prioritized automation backlog within hours of running the analysis.
- Understand the gap between OOTB agent coverage and your specific automation needs.
- Generate evidence for stakeholder reporting grounded in your own incident data.
Ongoing Adoption
Use AI Agent Advisor as a recurring health check to surface new automation opportunities as your data grows or as new agents are released.
- Run quarterly or after major releases.
- Identify drift: new issue patterns not covered by currently deployed agents.
- Feed output into AI adoption plans and executive review decks.
3. How Should I Interpret the Results?
Understanding the Dashboard
After the job completes, navigate to the AI Agent Advisor Dashboard. The results are organized into three panels:
Panel 1: Automation Opportunities
A table listing all identified clusters, sorted by the number of records that opportunity represents in descending order. Each row represents a distinct category of recurring issue in your operational data.
| Column | What it means |
|---|---|
| Name | An AI-generated label summarizing the cluster (e.g., "Active Directory password reset due to expiration"). |
| Available AI Assets | Number of AI Agents, Workflows, and Tools matched to that opportunity. |
| Est. ROI per year | Dollar value estimate of the Return On Investment if automated. Value stems from a calculation with configurable inputs. |
| Time savings per month | Estimate of the time saved per month for your fulfiller teams if automated.Value stems from a calculation with configurable inputs. |
| Number of requests | Number of incidents in the cluster. Higher values indicate higher potential automation impact. |
| Last analyzed | The date the analysis was performed. |
| Type | The type of record that was used to generate this opportunity. |
Panel 2: Resolution Steps and Matches
Click any opportunity to enter a more detailed view. This page consists of: Each step For each step, you will see:
- A summary of the issue, with values populated for:
- Number of requests: The number of requests included within this cluster
- ROI: Estimate of the return on investment if fully automated
- Hours saved: Estimate of the time saved if fully automated
- Resolution steps: The number of resolution steps typically taken to resolve that class of issue
- Ready to use AI solutions: Number of AI Agents, Workflows, and Tools mapped to potentially automate steps in the resolution process
- A series of resolution steps representing an action typically taken to resolve that class of issue. With these steps, you may find:
- The matched OOTB AI Agent name and Tool. (If present)
- A confidence score indicating how well the agent or tool fits the step.
- An explanation describing why this match was selected.
Panel 3: Sample Records
A sample of the actual records grouped into the selected cluster. Use this to validate that the cluster reflects a real, recognizable pattern before taking action.
Best Practices
Early implementation across multiple environments produced the following indicative findings:
- The best results occur within larger datasets, as trends are become more evident.
- Focus on the top 5-10 opportunities by record volume for the highest immediate ROI.
4. Additional Details
Licensing
| Attribute | Detail |
|---|---|
| Required Package | Active license for any Now Assist package |
| Activation | Enabled by default when any Now Assist plugin is installed. Can be toggled off from within the configuration options. |
| Pricing | AI Agent Advisor does not consume any assists. |
