AI Studio is available for Starter, Advanced, Enterprise, and Enterprise+ customers. Visit asana.com/ai to learn more about the launch of Asana AI Studio.
This article will help you troubleshoot common AI Studio issues and keep your Smart workflows running smoothly. To deepen your skills, you can also earn the AI Studio Foundations skill badge in the Asana Academy, where you’ll learn how to write effective prompts, plan workflows, and handle errors with confidence.
If you @mention multiple projects, the system will include as many tasks from those projects as can fit within the context window, up to a maximum of 200 tasks. The included tasks are sorted by last modified date. So if the data exceeds the limit, some tasks from the @mentioned projects may not be sent to Asana AI.
This usually indicates Asana AI does not have sufficient context to take the requested action clearly. Viewing Asana AI’s reasoning will explain the specific issue and suggest how to resolve it. You can learn how to do this in the “Check Asana AI's reasoning” section below. Often the solution is providing additional instructions so Asana AI has the necessary information to proceed confidently. Sometimes the instructions given to Asana AI need to provide additional context.
Asana AI’s reasoning gives a valuable insight into how Asana AI is processing information. Check Asana AI’s reasoning in each task’s history to understand how AI is interpreting your instructions.
The reasoning can show you things like:
To find Asana AI’s reasoning, open a task where Asana AI has taken action, scroll down to the task history/comments section and click Show reasoning as shown in the gif above.
To ensure your workflow functions as intended, it's important to set up your Smart workflow so it can perform the desired action. A common mistake is assuming that telling Asana AI what action you want it to perform in the instructions is enough, but remember that AI Studio is built on top of Asana's rules engine, meaning you will still need to create an action when building the rule before anything can happen.
For example, you might write instructions for Asana AI to comment on incoming requests that don't have enough information, asking the requester specific follow up questions. While these instructions explain how you want the action executed, Asana AI can't carry them out unless you also create an Add Comment action in the rule. Additionally, you must include the Use AI variable in the comment field when creating the action for the process to be completed successfully. Without these steps, Asana AI won't be able to follow through, no matter how clear your instructions are.
Think of the instructions as guidance for Asana AI on how to perform the action. However, creating the action in the rule and adding the Use AI variable specifies what the rule needs to do. Both elements are essential to make the workflow function smoothly.
An issue with attachments may be due to one of the following:
| Attachment type | Direct upload | Google Drive | OneDrive** | SharePoint** | Dropbox | Box |
|
Cloud folder (Google docs, etc.) |
n/a |
✅ |
n/a |
n/a |
n/a |
n/a |
|
Cloud doc (Google docs, etc.) |
n/a |
✅ |
n/a |
n/a |
n/a |
n/a |
|
Cloud sheet (Google sheets, etc.) |
n/a |
✅ |
n/a |
n/a |
n/a |
n/a |
|
Cloud presentation (Google slides, etc.) |
n/a |
✅ |
n/a |
n/a |
n/a |
n/a |
| Images (.jpg, .png, .apng, .tiff, .svg, .bmp, .gif, .webp, etc.)* | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| .docx | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| .xlsx/.csv (spreadsheets) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| .pptx (Microsoft PowerPoint slideshow) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | |
| .txt | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| .vtt (video text tracks) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| .json / .yaml | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
*Images can be read only by Anthropic's Claude models, not OpenAI's GPT models.
**If you disconnect a OneDrive or SharePoint account from Settings > Apps, you may still remain signed in on the Microsoft side. To fully switch to a different account, you may need to clear your browser cookies to remove the previous Microsoft login session.
Note
Dropbox is not currently supported.
You must explicitly mention other projects or tasks in order for Asana AI to get access to them. If Asana AI is not accessing the work that you expect:
See our article on Connecting your Smart workflow to the right data sources for more detailed information and tips to ensure Asana AI can access the work it needs.
If your Smart workflow is unable to see parent and subtask information, there are a few key settings to check. First, ensure you have chosen Task data, comments, activity, linked tasks under Data to use from tasks in the Data and permissions section of the instructions. If you've already enabled this setting but still experience issues, try explicitly @mentioning the parent or subtask in your workflow instructions. It's important to understand that these relationships aren't automatically visible to Asana AI without the proper permissions being in place.
Smart workflows don't automatically have visibility to all tasks within the same project, even when working with related items. To ensure proper access, you should always explicitly @mention the specific project in your workflow instructions. It's also crucial to verify your project data access settings are properly configured - these can be set to Project names only, Similar tasks, or Recent tasks. Remember that simply being in the same project doesn't automatically grant Asana AI access to other tasks.
When working with a project’s tasks in Smart workflows, the level of access is determined by your Data to use from projects setting. This setting has three options that provide different levels of visibility: Project names only provides minimal access and only allows the workflow to see project names; Similar tasks grants access to up to 10 tasks that are similar to the one being processed; and Recent tasks provides the broadest access, allowing visibility of up to 200 recently modified tasks. Understanding these settings is crucial for ensuring your workflow has the appropriate level of access to tasks within a project.
If your Smart workflow is having trouble adding collaborators, you'll need to check the Who can be added as a collaborator setting.
In the example below, the setting chosen only allows those @mentioned in the instructions to be added as collaborators when the rule creates a comment on the task.
This setting has three permission levels:
Selecting the appropriate setting will ensure your workflow can add the intended collaborators while maintaining your desired security level.