Asana has been steadily integrating AI features our work easier.
If you are looking to get the most benefit from those features and become an Asana AI Expert, this topic is for you!
0. Foundational
AI requires accurate information to function effectively. If relevant work communication isn’t in Asana, AI won’t be able to interpret it properly. (And neither will your colleagues.)
It’s a good habit anyway to name things well and keep the communication about work up to date in Asana. Aim to be so clear, so that you would be able to re-assign anything (tasks/projects/portfolios) without necessarily planning a meeting to get your colleague up to speed.
1. Where to start?
Asana’s Help Center has got you covered there with Get started with Asana AI. Here you’ll find an overview of all the features. If you want you can click through to specific articles for details, but for now the summaries should do.
Asana AI features need to be enabled in the Admin Console. For more guidance check out the Asana AI features and admin controls Help Center article.
Once AI features are enabled, I recommend starting your learning journey with the AI for Work Skill badge. A short course designed by Asana to give you a good grasp of the basics. From a basic understanding of what AI is to how to work with the different features.
Completing this course requires taking a small test (20 questions), and you can add the certificate to your LinkedIn or otherwise reference it.
2. I’ve got the badge and learned the basics, now what?
Now practice. Everywhere you see such a button in Asana: 
, press it and give it a spin!
Having learned the basics should save you some time, but taking full advantage requires deliberate practice.
Do experiments, switch up your prompts, and try some different angles to see what gives you the best output.
3. AI Studio overview
Now all the other AI features can definitely save you some time, but AI Studio really is the game changer.
In short, AI Studio allows you to integrate AI into your Asana rules as:
- conditions: AI decides whether a statement is true.
- actions: AI decides how to best execute the action. For example: what fiel values to apply, how to rename a task, what comment to make, etc…
The best thing is, you can mix and match it with non-AI rules!
4. Get AI Studio
AI Studio isn’t included in any plan by default, so to learn more about obtaining AI Studio check out this Help Center article: AI Studio add-on and pricing
5. Learn how to use AI Studio course
Again, our firends at the Asana Academy have made us a course: Building Smart workflows with AI Studio
Another great way to learn about AI studio is to join the AI Studio forum category. Here you can get inspiration and ask questions specifically related to AI Studio.
6. Practice some more
Again, to master something you need to practice. Experiment with different combinations, and keep refining.
After AI Studio has executed an action, you will be able to click “Show reasoning” in the task activity to better understand why AI Studio did what it did.
As AI Studio uses credits per condition/action executed it is good not to overdo it. So prevent using AI where it isn’t needed, and be critical whether you really need those heavier models.
7. A word of caution
Keep a human in the loop. While it might be an excellent idea to let AI propose a response to an incoming query, I would question the wisdom of sending out the reply unchecked. As LLM’s are predictive models, they can be biased to what they see most often. And the fact that something occurs often doesn’t mean that it is correct.
Extra caution would be advised when connecting AI to external communication for two reasons:
- You/your company might be held accountable for what it communicates
- It might pose a security risk, as these models might be tricked.
7. Advanced
If you really want to make the best functioning AI workflows I’d advise to also research the different models and their fitness for your purpose.
A technique that might be worth exploring would be to alternate AI and human interaction in the process.
For example:
- Article ideas are submitted through a form that asks a subject, target audience and purpose.
- AI drafts a rough outline for an article based on the form submission and moves it to the “refine outline” section
- Someone refines the outline to improve the quality and moves it to the “draft article” section
- AI drafts the article based on the refined outline and moves it to the “refine article” section
- Someone refines the article.
Where usually the interactions with LLM’s can be quite hit and mis, these steps in between give opportunity to steer the process and improve the quality of the outcome.
8. Use cases
Here I’ll share some ideas of how to apply AI Studio
- Use it as a Sales assistant (Video by @Kelsea_Lopez)
- Automating quality check for incoming briefs briefs and Check for duplicate tasks (Workshop by @Arthur_BEGOU and myself)
- HR assistant: Attach your HR manual and let AI studio draft replies to queries to HR.
- Translator: Facilitate people working in their native language by having AI Studio translate. For example, you’re in a team that has both Dutch and English speaking team members. Now you cana add a “translate to Dutch” and a “translate to English” button using manual rule triggers, and have AI replace or append the description.
The possibilities are near endless.
What will you build?
Jan-Rienk - Asana Expert @ Improving Every Day
