web
You’re offline. This is a read only version of the page.
close
Skip to main content

Announcements

News and Announcements icon
Community site session details

Community site session details

Session Id :
Power Platform Community / Forums / Copilot Studio / Best Practise for deve...
Copilot Studio
Suggested Answer

Best Practise for developing and deploying agents

(1) ShareShare
ReportReport
Posted on by 85

What is the best way, what is best practice when it comes to developming, versioning and deploying agents in Copilot Studio?

 

I now basically have three ways, all of which do the same thing but deliver different data structures:

 

1. I can export the agent's solution manually from my environment via the solution's website and have the version number incremented automatically.

2. I can use the PAC CLI tools to check out a solution locally, edit it, and check it back in. However, I have to delete the folder before each checkout. This feels strange—so different from working with git, for example.

3. I can also check out an agent using the Copilot Studio Extension in VS Code, edit it, and then apply the changes to the cloud.

 

The files and folders from solutions 1 and 2 look almost identical—though not 100%. They are XML files for the bot components and separately stored content in data.

In case 3 with VS Code, however, I have yml files that are also stored in a completely different folder structure.

 

What is the best way to proceed at the moment? And does it make sense to edit agents both via the web interface in Copilot Studio and via the local files (always bearing in mind, of course, that you first update/check out everything before working locally)?

 

Or is there another way that I am not yet aware of?

I have the same question (0)
  • Suggested answer
    Beyond The Platforms Profile Picture
    118 on at
    Hi,
    You're not alone 😊
    Here's the current recommended approach:
    Use both tools, for different purposes:
     
    • VS Code extension (YAML) → day-to-day development, local editing, Git/PR workflow. This is the closest thing to a proper developer experience and the direction Microsoft is pushing.
    • PAC CLI / Solution export (XML) → environment promotion (Dev → Test → Prod), CI/CD pipelines via Azure DevOps or GitHub Actions.
    The two outputs (YAML vs XML) are intentionally different: YAML is for authoring, XML is for solution transport. Don't try to mix them.
    About the PAC CLI "delete before checkout" issue — that's by design. PAC CLI is not meant to replace Git, it's a transport tool. The VS Code extension is what fills the Git-like gap.

    Can you mix web UI and local edits? Yes, but always pull from cloud before editing locally and avoid parallel edits in both places at the same time.

    A fourth option worth knowing: Power Platform Pipelines — a built-in, click-based promotion tool that doesn't require setting up Azure DevOps or GitHub Actions. Great if you want governance without heavy DevOps infrastructure. I don't know if you've ever tried it
     
    Hope this helps!
    Paolo
     
    Did this solve your issue? → Accept as Solution
    👍 Partially helpful? → Click "Yes" on "Was this reply helpful?" or drop a Like!
     
    Want more tips on Power Platform & AI? Follow me here:
     
  • Konrad Profile Picture
    85 on at
    Thank you for your detailed reply :-)
     
    But editing the yaml-files in VS Code and then directly uploading them to "the cloud" is only recommended for the test instances of an agent, isn't it?

    Editing YAML --> upload to test agent --> and then ... doing the same (editing yaml) for the production agent? Or do the editing of the once tested changes in the PAC/CLI checkout and packing/import it to the agent in production only via PAC?
  • Suggested answer
    Beyond The Platforms Profile Picture
    118 on at

    Hi Konrad,

    The recommended approach is:

    1. Edit YAML via VS Code → push/apply to your Dev environment (or a dedicated test agent). This is your authoring layer.

    2. Once tested and validated, export the solution from Dev as a managed solution (via PAC CLI or the web UI — the XML-based solution package).

    3. Import the managed solution into Test, and then Production, using PAC CLI or Power Platform Pipelines.
    Try this.
     
    Hope this helps!
    Paolo


    Did this solve your issue? → Accept as Solution
    👍 Partially helpful? → Click "Yes" on "Was this reply helpful?" or drop a Like!


    Want more tips on Power Platform & AI? Follow me here:

    🔗 LinkedIn: https://www.linkedin.com/in/paoloasnaghi/
    ▶️ YouTube: https://www.youtube.com/@BeyondThePlatforms
    📸 Instagram: https://www.instagram.com/beyond_the_platforms/
    🌐 Website: https://www.beyondtheplatforms.com/


     

Under review

Thank you for your reply! To ensure a great experience for everyone, your content is awaiting approval by our Community Managers. Please check back later.

Helpful resources

Quick Links

Introducing the 2026 Season 1 community Super Users

Congratulations to our 2026 Super Users!

Kudos to our 2025 Community Spotlight Honorees

Congratulations to our 2025 community superstars!

Congratulations to the March Top 10 Community Leaders!

These are the community rock stars!

Leaderboard > Copilot Studio

#1
Valantis Profile Picture

Valantis 586

#2
chiaraalina Profile Picture

chiaraalina 170 Super User 2026 Season 1

#3
deepakmehta13a Profile Picture

deepakmehta13a 118

Last 30 days Overall leaderboard