Loading

Topics

Optimize AI credit usage in automations

Who can use this feature?
StarterAdvancedEnterpriseEnterprise+PremiumBusinessLegacy Enterprise

Visit our pricing page for more information.

Automations with AI steps use AI Studio credits each time they run, so the way you design an automation affects how quickly you use them. This article covers four common design patterns that lead to higher credit usage than expected, and simpler alternatives for each. 

Related articles

How credits are consumed

Each time an automation with AI steps runs, it consumes credits. The amount depends on how much information the automation reads, how much it writes, which AI model it uses, and how often it runs. Admins can track credit usage in the admin console. Design for the best fit first, then look for ways to reduce credit usage without lowering the quality of the output.

Limit how often automations run

Automations that run every time a task is created, added to a project, or updated can run more often than you need, especially in busy projects or on frequently updated fields such as status or progress. Add conditions so the automation only runs when it will be most useful. 

Run automations item by item only when timing matters, such as assigning a task as soon as it is created or alerting someone the moment a deadline is missed. For reporting, summaries, and reviews, ask whether running less often would give the same result.

Admins can also set an hourly credit limit that pauses additional runs when an automation reaches it, which helps protect against sudden spikes in credit use. Asana provides a default limit that admins can change, and the same limit applies to all automations; it is not set per automation or per user. For more information, see AI Studio rate limiting.

Keep AI steps focused

Automations use more credits when they do more. Capabilities such as web search, attachments, and premium models can add to usage without necessarily improving the result. A very large automation can use significantly more credits per run and may be less reliable, so keep instructions concise and avoid repeated context. Shorter, cleaner instructions often produce better output at lower usage.

Use AI conditions only when they add value

Automations that use AI conditions use credits each time they are triggered, even when no action follows. The cost of a single run is small, but it adds up quickly when paired with a trigger that runs often. For example, an automation that uses AI to check every comment on a task for a risk will run, and use credits, on every comment.

Limit the data an automation reads

Automations that reference large portfolios or entire projects read far more data than most workflows need. This raises usage and can lower the quality of the output. Scope the automation to only the projects that matter, and use a custom field or tag to filter what it needs to see.

 

Loading
Optimize AI credit usage in automations | Asana Help Center