Configuring AI Tools Surveys, and Application Vocabulary in Proactive IT Experience
AI Tools surveys help organizations measure employee experiences with AI solutions, including Microsoft Copilot, ChatGPT, Claude, and internal AI assistants. Using the same application vocabulary framework as Enterprise Applications surveys, AI Tools surveys provide a consistent way to categorize tools, target users, and analyze AI adoption and employee experience across the organization.
How AI Tools Surveys Work
AI Tools is a survey type within Proactive IT Experience. Each survey measures employee experience for one or more AI tools defined in the application vocabulary.
An AI Tools survey consists of the following components:
- Main Item
- Alternative Items
- Root Category
- Source Record
- Target Audience
Main Item
The primary AI tool being evaluated.
Examples:
- Copilot
- ChatGPT
- Claude
- Gemini
Alternative Items
Optional additional AI tools that respondents can select when answering the survey.
Root Category
A vocabulary category used for organizing applications and reporting.
Source Record
An optional link to an application, business application, CMDB record, or another ServiceNow record.
Target Audience
Determines which users are eligible to receive the survey.
Example vocabulary structure:
AI
- Copilot
- ChatGPT
- Claude
- Gemini
Creating AI Tool Vocabulary Items
When creating an AI Tools survey, the Main Item must reference a vocabulary item that represents the AI solution being measured.
If a suitable item does not already exist, create a new vocabulary item.
Recommended vocabulary names
✔ Copilot
✔ ChatGPT
✔ Claude
✔ Gemini
✔ Internal AI Assistant
Avoid technical system names
✘ GPT4-WestEU-01
✘ AIPROD-INSTANCE-3
✘ M365-COPILOT-TENANT
Users should immediately recognize the tool they are being asked about without needing technical knowledge.
Choosing a Root Category
The Root Category is used for classification and reporting.
Categories do not need to mirror your CMDB or ServiceNow application structure.
Example:
IT
- Collaboration
- Productivity
- Access
- AI
For most organizations, a simple structure is sufficient:
IT
- AI
Keeping the category structure simple generally improves reporting and administration.
Single AI Tool vs Multiple AI Tools
Single AI Tool Survey
Use this approach when measuring one specific AI solution.
Example:
Main Item:
- Copilot
Alternative Items:
- None
Recommended notification:
- Happy IT AI Tool
Multiple AI Tools Survey
Use this approach when users have access to several AI solutions.
Example:
Main Item:
- Copilot
Alternative Items:
- ChatGPT
- Claude
- Gemini
Recommended notification:
- Happy IT Multiple AI Tools
In this configuration, respondents can choose which AI tool they want to evaluate.
Linking AI Tools to ServiceNow Records
Vocabulary items can optionally be linked to ServiceNow records.
Examples include:
- Business Applications
- CMDB Application records
- Custom application tables
This configuration is completely optional.
The survey will function without any record linkage.
However, linking vocabulary items to application records can improve:
- Reporting
- Filtering
- Application administration
- Future integrations
Example:
| Vocabulary Item | Linked Record |
|---|---|
| Copilot | Business Application |
| ChatGPT | CMDB Application |
| Internal AI Assistant | Custom Application Record |
Best Practice: Use Dynamic Target Audiences
The most effective targeting model is to use groups that represent users of the AI solution.
Recommended examples:
- Entra ID groups
- Active Directory groups
- Security groups
- Software access groups
Example:
Copilot Licensed Users
As users are added to or removed from the group, survey eligibility updates automatically.
Recommended
Use groups that represent actual users of the tool.
Not Recommended
Avoid:
- Static user lists
- Manually maintained audiences
- Individual user selections
Dynamic group-based targeting reduces administration effort and ensures surveys are sent only to relevant users.
Using Source Tables for Dynamic Audiences
Target audience source tables must contain a relationship to a user record.
Examples include:
- User tables
- Group membership tables
- Application access tables
- Security group membership tables
A common implementation pattern is:
- Users receive access to an AI tool through a security group.
- The security group is synchronized into ServiceNow.
- The survey audience is built from that group.
- Survey recipients are calculated dynamically each time the survey process runs.
This allows audience membership to remain current without manual maintenance.
AI Employee Questions
AI Tools surveys can be combined with Employee Data questions to provide additional context.
Examples include:
- AI confidence
- AI skill level
- AI usage frequency
- Technology mindset
These responses can later be used as filters in reporting.
Example use cases:
- How do employees with high AI confidence rate Copilot?
- How do less experienced AI users rate ChatGPT?
- Do frequent AI users report a better overall IT experience?
Combining AI Tool surveys with Employee Data questions often provides significantly deeper insights than measuring the tool experience alone.
Notifications
Two out-of-the-box notifications are available for AI Tools surveys.
Happy IT AI Tool
Used when surveying a single AI tool.
Happy IT Multiple AI Tools
Used when respondents can select from multiple AI tools.
Organizations can clone and customize these notifications to align with their branding, terminology, and communication style.
Supporting Multiple AI Tools
Some organizations prefer to measure overall AI adoption instead of evaluating only one solution.
In this scenario, a survey might use:
Main Item:
- Copilot
Alternative Items:
- ChatGPT
- Claude
- Gemini
- Internal AI Assistant
This approach allows a single survey campaign to collect feedback across multiple tools while maintaining tool-specific reporting.
Categories and Reporting
Categories can be used to organize AI solutions into logical groups.
Examples:
IT
- AI
or
IT
- Productivity AI
- Copilot
- ChatGPT
- Gemini
- Development AI
- GitHub Copilot
- Cursor
- Internal AI
- Company AI Assistant
Choose a structure that matches how your organization reports and discusses AI adoption.
Language and Time Zone Support
Time Zones
Survey delivery follows the user's ServiceNow time zone whenever available.
The delivery order is:
- User time zone
- Survey default time zone
- ServiceNow system time zone
Using the recipient's time zone is recommended, especially for global organizations.
Language
Survey language is determined by the user's preferred language setting in ServiceNow.
If a translation is available, the survey will automatically be displayed in the user's preferred language.
Note that notification template translations may need to be configured separately in ServiceNow.
Recommended AI Survey Model
For organizations starting AI experience measurement, the following model is recommended.
Vocabulary
Root Category:
- AI
Main Item:
- Copilot
Alternative Items:
- ChatGPT
- Claude
- Gemini
Target Audience
- Entra ID group containing licensed Copilot users
Notifications
- Happy IT Multiple AI Tools
Additional Employee Data Questions
- AI confidence
- AI skill level
- AI usage
- Technology mindset
This approach provides visibility into both the employee experience of AI tools and the overall maturity of AI adoption within the organization.
Best Practices Summary
- Use recognizable AI tool names that employees already know.
- Keep the vocabulary structure simple.
- Use group-based audience targeting whenever possible.
- Link vocabulary items to ServiceNow applications if reporting benefits from the relationship.
- Enable AI-related Employee Data questions to enrich reporting.
- Use multiple AI tools in a single survey only when users genuinely have access to multiple solutions.
- Start with a pilot audience before expanding to the wider organization.
- Review survey results together with AI confidence and AI usage data to better understand adoption challenges and opportunities.