Tracking Now Assist Usage: Measuring ROI and Adoption (Zurich)

[Dexter Chan]

Tracking Now Assist Usage: Your Complete Guide to Measuring ROI and Adoption 🎯 

If you are responsible for managing Now Assist in your ServiceNow environment, you have likely asked yourself these questions: Which teams are using Now Assist the most? How many assists/tokens are we consuming? Are we getting real value from our investment? Understanding usage patterns is not just about monitoring consumption. It is about proving ROI, optimizing deployment, and making data-driven decisions to refine your AI strategy.

The challenge is that many platform owners are unsure where to look for this information. In this guide, we will walk through everything you need to know about tracking Now Assist usage in ServiceNow.

What You Will Learn

By the end of this post, you will be able to:

1. Understand the difference between "actions" and "assists" (and why it matters for billing)

2. Navigate to Now Assist Analytics and understand the out-of-box dashboards available to you

3. Interpret usage data across different areas including departments, skills, and channels

4. Understand how subscription management provides visibility across all instances

Assists vs Actions 101

Now Assist skills are generative AI skills designed to meet the needs of users in different workflows. Each Now Assist Skill is comprised of one or more “actions” based on the complexity of the skill. An "assist" is ServiceNow’s version of a token that measures usage of a Now Assist skill.

Different skills consume different amounts of assists. A straight-forward skill (ex. case summarization) is comprised of few actions so it consumes one assist and a complex agentic workflow is comprised of multiple actions and can consume 25 assists or more. See the entitlements package for full details on the assists. Overall, each customer has a number of available assists defined by their contract, and their sales account teams may be consulted for how to best optimize these assists.

Finding Now Assist Analytics

Now Assist Analytics is your centralized location for monitoring AI usage in your instance. This is a store application that requires no additional licensing beyond your Now Assist subscription. It provides dashboards that help you understand feature adoption, skill performance, and user behavior.

To access Now Assist Analytics, navigate to All > Now Assist Admin > Analytics. This takes you to a collection of purpose-built dashboards, each designed to answer specific questions about your Now Assist journey.

Note: these dashboards have been updated frequently since their initial release. If you are on Xanadu or newer releases, make sure you have installed the latest version from the ServiceNow Store to access all available insights.

Understanding the Out-of-Box Dashboards

Usage and Adoption Dashboard

This dashboard gives you a clear snapshot of how people are actually using Now Assist. You can filter the dashboard by date range, channel (now assist panel, virtual agent, in product), and product family (ITSM, CSM, HRSD) to drill in on the slices of usage that matter to you.

Key usage summary metrics include:

• Total Now Assist actions: Total number of actions used

• Daily Now Assist actions: Total number of actions used in a day

• Average daily unique users engaging with Now Assist: Helps you see whether engagement is growing or flattening out

• Daily unique users engaging with Now Assist: Day-by-day view of active users

• Skill group distribution: Shows where Now Assist is delivering value across different skill categories. This chart can have the data broken down by feature, family, skill, and channel

• Daily usage comparison by workflow: Helps you understand which workflows drive the most Now Assist engagement

• Skill engagement trends: Reveals which skills users are interacting with most over time

• Adoption by department: The Adoption tab shows which departments have the highest usage and which skills they use, highlighting adoption patterns across your organization

• Feedback details and Error details: Quick signals about quality and reliability feedback. This shows how users respond to the responses they are getting, while error patterns help you spot configuration or data issues before they become bigger problems

Key adoption metrics include:

• Departments with highest usage: Shows which departments have the highest Now Assist action usage in your selected period of time

• Now Assist actions comparison by user department: Further breaks down the top five departments and which skills are used the most 

• Feedback details and Error details: Quick signals about quality and reliability feedback. This shows how users respond to the responses they are getting (ie thumbs up or down), while error details help you identify which Now Assist skill family are returning the most answers with an error status

Together, these metrics give developers and admins a practical way to track adoption and understand where Now Assist is making the biggest impact.

Self-Service Performance Dashboard is no longer visible and has been retired in Zurich - use Assistant Analytics Hub

Skill Performance Dashboards

Now Assist Analytics provides three dashboards for tracking skill performance at different levels of detail. Start with the high-level Skill Performance dashboard to identify trends, then drill into Skill Details or Custom Skill Details to investigate specific capabilities.

Skill Performance Dashboard

This is your high-level view across all Now Assist skills in your instance. Use this dashboard to identify which skills are gaining traction and where you should focus your attention.

Key metrics include:

• Skill engagement trend: Visualize adoption patterns across all skills over time

• Number of actions: Total volume of skill usage

• Total daily active users: How many unique users are engaging with Now Assist each day

• Total daily active users by skills: Break down daily engagement by individual skill to spot which capabilities drive the most activity

Skill Details Dashboard

This dashboard focuses on out-of-box Now Assist skills provided by ServiceNow. At the top title, there is a picker to select a Now Assist skill that you would like to dive deeper into. 

Key metrics include:

• Skill engagement trend: Track usage over time for each out-of-box skill

• Total skill actions: Total number of times the skill has been used

• Accepted skill actions: How many skill executions users accepted or acted upon

• Acceptance rate: Percentage of skill actions that users found valuable (ie accepted)

• Daily active users: Number of unique users engaging with the skill each day

Custom Skill Details Dashboard

This dashboard provides visibility into the Now Assist skills you have built or customized in a defined time period. It helps you prove the value of tailored AI solutions with metrics that show how your custom skills perform across different contexts.

Key metrics include:

• Number of custom skills created:  The total number of custom skills created to date in the instance

• Number of custom skills activated: The total number of custom skills that are in active state

• Number of prompts in active custom skills: How many prompts your active skills are running. More prompts lead to more complexity to manage, so tracking this helps you spot when things might be resource intense

• Number of Assists consumed by custom skills: The total number of assists that have been used to execute your custom skills based on the defined time period selected. Note: the number in green/red will show you a baseline count from the last time you selected this time period

• Count of invocations: The total number of actions used in the defined time period 

• Average number of unique users using custom skills: The average number of users who use each of the custom skills based on the defined time period

• Daily unique users engaging with the skill: Track daily adoption of your custom capabilities

• Skill engagement trend by workflows: See how your custom skill performs across different business processes

• Skill engagement trend by products: Break down usage by product family (ITSM, CSM, HRSD, etc.)

• Skill engagement trend by features: Understand which specific features drive engagement

• Executed successfully: Monitor the technical success rate of your custom skills (ie accepted executions with no error)

• Skills feedback: Capture user sentiment about your custom solutions

Important note: If you are on a pre-July 2025 version, you may only see three dashboards. The enhanced skill detail dashboards require an update to the Now Assist Analytics application.

Now Assist Guardian Dashboard

Offensive Content Dashboard

For organizations using Now Assist Guardian to filter offensive or inappropriate content and prevent prompt injection attacks, this dashboard provides visibility into how often Guardian is triggered and the impact on system performance.

Offensive Content Key metrics include:

• Guardrail-added latency: Measures the performance impact of Guardian's security checks on response times

• Percentage flagged as offensive: Shows what proportion of total interactions were identified as offensive or inappropriate

• Total offensive content occurences: Number of times Guardian detected and blocked problematic content

• Categories of offensive content: Breaks down the types of content being flagged (hate speech, harassment, etc.)

• Offensive content occurences by skill: Shows where offensive content is appearing across different Now Assist capabilities

This dashboard helps you understand both the effectiveness of your content filters and potential areas where users might be encountering friction. If you see high latency in certain skills, you can evaluate whether Guardian configuration needs adjustment. The occurrence data also reveals whether specific skills are more vulnerable to misuse, helping you prioritize security hardening efforts.

Prompt Injection Dashboard

The Prompt Injection sub-dashboard provides focused visibility into security threats where users attempt to manipulate Now Assist behavior through malicious prompts.

Prompt Injection Key metrics include:

• Guardrail-added latency: Performance impact specifically from prompt injection detection

• Percentage flagged as prompt injection: What proportion of interactions were identified as potential prompt injection attempts

• Total prompt injection occurrences: Overall count of detected prompt injection attempts across your instance

• Prompt injection occurrences by skill: Shows which Now Assist capabilities are most frequently targeted by prompt injection attempts

This focused view helps security and platform teams identify patterns in prompt injection attacks, understand which skills require additional hardening, and monitor whether malicious activity is increasing or decreasing over time.

User Search Analyzer Dashboard

This is one of the most actionable dashboards for continuous improvement. It shows you what users are searching for, which queries return AI-generated results, and critically, which searches come up empty.

Key metrics include:

• Queries with genius results: How often AI-generated answers successfully respond to user queries

• Genius Result engagement: Tracks whether a user has clicked to follow up or to display the full answer to an answer

• Genius Results response time: Measure of how quickly a response has loaded

• Genius Result to chat handoff: Number of virtual agent conversations that utilized ask a follow up

• Genius Result citation engagement: Number of virtual agent conversations that utilized multi-content synthesized genius results

• Feedback (Thumbs up): Number of positive feedback provided by users

• Feedback (Thumbs down): Number of negative feedback provided by users

• Genius Results with KB: Number of responses with a knowledge article provided as a source

• Genius Results with Catalog Item: Number of responses with a catalog item provided as a source

• Top 10 queries with KB genius results: Top ten search queries that resulted with a knowledge article provided as a source

• Top 10 queries with catalog item genius results: Number of responses with a catalog item provided as a source

Use this data to identify gaps in your knowledge base and prioritize content creation efforts.

Now Assist Context Menu Dashboard

The context menu is how users access Now Assist directly within their workflow whether it be right-clicking on form fields or text to invoke AI capabilities. This dashboard tracks how often users discover and engage with this interaction pattern.

Key metrics include:

• Usage in this period: Total volume of context menu invocations in your selected timeframe

• Acceptance distribution: Shows how often users accept or act upon the AI suggestions triggered through the context menu

• Usage trend by skill: Tracks which Now Assist skills users access most frequently through the context menu over time

• Capability distribution: Breaks down context menu usage across different AI capabilities

• Responses by feedback: Captures user sentiment about context menu interactions (helpful, not helpful, etc.)

These insights help you understand whether users are discovering the context menu feature and finding it valuable in their daily work. If context menu usage is low despite having capable skills available, it may indicate a need for additional training or in-app guidance to help users discover this capability. The feedback data also reveals which context-initiated actions provide the most value, helping you prioritize which skills to promote through this interaction pattern.

Subscription Management: The Cross-Instance View

While Now Assist Analytics shows you what is happening within a single instance (usually a production instance), Subscription Management provides the enterprise-wide perspective. This is where you see total License and Now Assist across all your instances, including production, sub-production environments, and development instances.

To access this, navigate to Admin > Subscription Management > Subscriptions > Now Assist usage.

At the top of this view, you will see three critical numbers:

1. Total Purchased: Your contracted pool of Now Assist assists

2. Total Used: Assists consumed across all instances

3. Available Assists: Your remaining pool

Below this summary, you will see a breakdown by instance showing which environments are consuming the most assists. This is invaluable for understanding where usage is happening. Perhaps your development team is consuming the most assists while building custom skills, or maybe one business unit's production instance is far outpacing others.

In the Subscriptions tab > Account level entitlements > Now Assist usage, you will see:

• Skill/Agent Name: What was used

• Actions: How many times each skill has been used

• Assist ratio: The relative cost for each skill

• Total assists: Consumption calculated by multiplying the number of actions by the assist ratio. This shows which skills consume the most entitlements

This breakdown helps you understand not just who is using Now Assist the most, but what they are using it for and how that translates to costs. You might discover that a heavily-used lightweight skill consumes fewer assists than a rarely-used complex agent.

Special notes: As of Feb 2026, Subscription Management usage is tracked as a 365 burn down model where usage now resets annually based on a customer's contract anniversary date. Previously, it was a 365 look back date where usage was calculated based on previous 365 days of activity. The goal of this change is to provide easier transparency into Now Assist entitlements in a more user friendly manner. Further information can be found here.

Use the following when determining whether to use Now Assist Analytics vs Subscription Management:

• If you need real-time data, recommend using Now Assist Analytics on the specific instance

• If you need enterprise-wide trending, use Subscription Management with the understanding that it reflects recent but not immediate usage

For further training on Subscription Management, additional resources can be found below:

1. Platform Academy: All About Subscription Management
2. Subscription Management Resources List

Pro Tips for Success

1. Double check your clone configuration settings: The usage log table should be excluded from cloning by default, but verify this in your clone settings. If usage data gets cloned from production to development, your analytics will show misleading information.

2. Establish baseline metrics early: Capture your usage patterns in the first 30 days of deployment so you can measure growth and adoption over time.

3. Set up regular review cycles: Schedule monthly reviews of your usage dashboards with stakeholders to identify trends, celebrate wins, and address underutilized capabilities.

4. Watch for unexpected consumption spikes: If Subscription Management shows a sudden increase in one instance, drill into Now Assist Analytics on that specific instance to understand what changed. It could indicate successful adoption or could reveal that someone accidentally left a development process running.

5. Correlate usage with outcomes: Track Now Assist adoption alongside your existing service management KPIs. Are departments with higher Now Assist usage also seeing improved ticket resolution times or higher customer satisfaction scores?

6. Create role-specific dashboards: Executives care about ROI and deflection rates. Service desk managers care about which fulfillers are adopting AI tools. Platform owners care about technical performance. Build targeted views for each audience rather than trying to make one dashboard serve everyone.

What's Next?

Once you have mastered usage tracking, consider these next steps:

1. Explore User Experience Analytics: This platform feature allows you to instrument custom events and track user journeys through your portals and workspaces, giving you even deeper insights into how Now Assist fits into your users' workflows.

2. Implement automated alerting: Use Performance Analytics indicators to trigger notifications when usage drops unexpectedly or when specific skills show declining adoption.

3. Build custom skills and track their performance: Now that you understand how to measure usage, you can confidently create organization-specific Now Assist skills and prove their value through data.

4. Attend the Platform Academy: ServiceNow hosts regular academy sessions focused on AI capabilities, including deep dives into configuration, optimization, and advanced use cases.