Deploying and monitoring the skill

After you evaluate the prompt, you can deploy the custom skill and monitor its effectiveness.

AI Skill Kit deployment and monitoring overview

You should build and evaluate the skill in a subproduction instance. After you’re satisfied with the final performance evaluation on a batch of test data in subproduction, you can begin to move a skill toward deployment.

1. Identify the touch points where the skills can be surfaced to be available for the end user. Configure the deployment settings accordingly.
2. Perform end-to-end user testing in subproduction for a while and with different users and inputs.
3. After the results are acceptable, deploy the skill to the production instance.
4. Monitor the skill.