How can we optimize and control assist consumption and why aren't assist unlimited?

[Suzanne4]

I am trying to understand how to configure the skills correctly to ensure we do not go over the number of assists we are licensed for.  Why is the model limiting in this way?  If I can understand why, it will help us to set things up correctly.

[VishalC3]

HI @Suzanne4 ,

The key thing to understand: an assist is "a unit to measure usage of Now Assist skills via performed skill actions" . So you're not licensed for a number of users or features. You're licensed for a pool of actions. The reason the model works this way is that the underlying LLM cost scales with how much work the AI does.

Let me know if you need more details and I can explain few ways to limit your usage. 

Thanks,

Vishal

[Vamsi_Krishna07]

Hello @Suzanne4 ,

The limit exists because each execution of a Now Assist capability consumes assists based on the type of AI feature being used. For example, a Now Assist Skill typically consumes fewer assists than an Agentic Workflow, which may involve multiple AI operations and therefore has a higher assist cost.

The goal of the configuration is to ensure AI capabilities are used where they provide the most value while staying within your licensed assist allocation. By enabling only the required skills, controlling where they are available, and monitoring assist consumption through analytics, you can effectively manage usage and avoid exceeding your licensed capacity.

How to track the assist consumption in your instance:

1. Navigate to Subscription Management -> Go to Now Assist Usage tab -> Look at the different reports present over there.
2. Navigate to Now Assist Admin Console -> Open Performance tab -> Look at the out of the box dashboards avaiable.

Thanks!