---
sourceDocument: Brazil Enable AI
sourceDocumentLink: https://www.servicenow.com/docs/r/intelligent-experiences

 Release :

    - brazil

ft:locale :

    - en-US

ft:publication_title :

    - Brazil Enable AI

ft:clusterId :

    - platai

bundleId :

    - platai

workflow :

    - Platform


---

# General guidelines for AI Skill Kit

# General guidelines for AI Skill Kit {#ariaid-title1}

Release version: Brazil  
Updated September 24, 2026  
![](https://www.servicenow.com/docs/portal-asset/ico-clock) 1 minute to read  
Building a custom skill in AI Skill Kit follows five phases, from scoping the skill to deploying and monitoring it.

## AI Skill Kit skill-building phases {#na-skill-kit-guidelines__gg-overview}

Think about the process of building a custom skill as having the following phases:

1. [Scoping the skill](https://www.servicenow.com/docs/~kcGe9B2euTn9PgD2jzwtg "Scope a custom skill before you build it to define its requirements and expected outcomes.")
2. [Collecting data and creating a dataset](https://www.servicenow.com/docs/yB22~ZdOvjEiuP2cvNTiPA "A representative, appropriately sized, and isolated dataset gives you a reliable measure of how your skill prompt performs.")
3. [Developing the prompt](https://www.servicenow.com/docs/NqCeETTClA_ilgxcLptEug "Write a skill prompt that is specific, clear, and grounded in context so that the model is more likely to return the output that you expect.")
4. [Conducting a performance evaluation](https://www.servicenow.com/docs/JdahrOrGTSMhBrihao8UtQ "Evaluate a skill prompt during development and again before deployment, using batches of test data and metrics that suit your use case.")
5. [Deploying and monitoring the skill](https://www.servicenow.com/docs/vmB4rldgD5AE4vBrsxopbA "After you evaluate the prompt, you can deploy the custom skill and monitor its effectiveness.")

Figure 1. Stages of building a custom skill

Adopt a data-driven mindset while you build a skill. A data-driven mindset means emphasizing evidence-based decisions and minimizing speculation about how a large language model (LLM) can interpret the specific wording of a
prompt.

Using AI Skill Kit requires intermediate prompt engineering skills. If you're not a developer, build that knowledge before you work with the tool.

Tooltips throughout the product explain individual fields and options.

To keep duties separate, a developer or generative AI practitioner creates the skill and then publishes it. After the skill is published, a AI Admin Hub activates the skill so that it appears at the configured touch points and is available to use.