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 / Power Automate / PAD Unattended: “Launc...
Power Automate
Suggested Answer

PAD Unattended: “Launch Outlook” fails with COMException 80080005 (CO_E_SERVER_EXEC_FAILURE)

(0) ShareShare
ReportReport
Posted on by 13

Hi all,

I’m looking for guidance on a recurring issue with Power Automate Desktop (PAD) automating Classic Outlook. The flow works reliably for a period (e.g., several runs across a day), but then starts failing again — usually the next morning / after a “cold start”.

Scenario


  • Type: Desktop flow triggered from cloud flow using “Run a flow built with Power Automate for desktop”

  • Run mode: Unattended

  • Machine: Hosted/VM (Windows) running PAD Machine Runtime

  • Licensing: The environment has the Hosted Process Machine license assigned (so unattended/hosted execution is enabled from a licensing perspective)

  • Outlook: Classic Outlook (M365 Apps) installed. New Outlook has been uninstalled.

  • Mailbox: Shared mailbox processing (mailbox access works in Outlook)


  •  

What the flow does


  • Launch Outlook

  • Read emails from a shared mailbox folder

  • Download attachments / process them

  • Move emails to completed subfolders (desktop actions)


  •  

The problem

In unattended runs, the flow sometimes fails at Launch Outlook with a COM exception.

It can work fine for multiple runs (e.g., 6 successful runs across a day), then fails again later (often after overnight / first run of the day).

Exact error

Cloud flow failure summary:

 

Action 'Run_a_flow_built_with_Power_Automate_for_desktop' failed: Problem while executing action 'Launch'. Failed to launch Outlook

PAD error details:
 

Microsoft.PowerPlatform.PowerAutomate.Desktop.Actions.SDK.ActionException: Failed to launch Outlook
---> System.Runtime.InteropServices.COMException: Retrieving the COM class factory for component with CLSID {0006F03A-0000-0000-C000-000000000046} failed due to the following error: 80080005 Server execution failed (Exception from HRESULT: 0x80080005 (CO_E_SERVER_EXEC_FAILURE)).
   at Microsoft.Flow.RPA.Desktop.Modules.Outlook.Actions.OutlookActions.LaunchOutlook(Variant& outlookInstance)

Interesting observations


  • If I switch the flow to Attended and watch it run on the VM, it generally completes successfully.

  • I changed the start logic to avoid PAD launching Outlook “from scratch” and instead:

    1. Run applicationOUTLOOK.EXE /recycle

    2. Wait a few seconds

    3. Loop / retry Launch Outlook several times until an instance is returne  


  •  

When I watched it in attended mode, I saw behaviour like:

 

  • Outlook UI starts successfully

  • Launch Outlook sometimes fails to pick up the instance at first

  • It may take 2–3 retries before it finally succeeds and the flow continues


  •  
 

This suggests a timing / cold-start / COM registration readiness issue rather than Outlook not being installed.

 

What I’ve already tried

 

  • ✅ Confirmed New Outlook is uninstalled and using classic Outlook

  • ✅ Confirmed Outlook opens normally when launched manually under the same service account

  • ✅ Added robust retries: start Outlook with /recycle then retry Launch Outlook

  • ✅ Ensured no conflicting Outlook instances: taskkill /F /IM OUTLOOK.EXE at start

  • ✅ Confirmed licensing/entitlement: Hosted Process Machine license assigned to the environment


  •  

My questions


  1. Is 80080005 (CO_E_SERVER_EXEC_FAILURE) in unattended PAD typically caused by Outlook COM not being ready (profile/auth/add-ins) even after OUTLOOK.EXE has started?

  2. Are there recommended best practices for stabilizing Launch Outlook in unattended runs? (e.g., longer waits, attach-only retries, warm-up flow, etc.)

  3. Is there a known issue where Launch Outlook fails intermittently after the machine is idle / after nightly session recycling?

  4. If you’ve solved this in enterprise environments, what changes made the biggest impact? (Outlook profile tweaks, disabling add-ins, cached mode settings, or specific PAD runtime settings)
I have the same question (0)
  • Suggested answer
    Haque Profile Picture
    3,653 on at
     
    It seems like Classic Outlook causing the issue. If there is no restrictions - can you test with the new outlook. Howerver, here are your answers, check if they can help.

    1. Is 80080005 (CO_E_SERVER_EXEC_FAILURE) in unattended PAD typically caused by Outlook COM not being ready (profile/auth/add-ins) even after OUTLOOK.EXE has started?

    Yes, this error often indicates that the Outlook COM server is not fully ready or accessible, even if the Outlook process is running. Common causes include:

    Outlook profile not fully loaded or authenticated. Add-ins or startup processes delaying COM server readiness. Outlook running but in a hung or unstable state. Permissions or session context issues in unattended mode.

    Note: Simply starting OUTLOOK.EXE does not guarantee the COM interface is ready for automation, which is why retries and waits are often necessary.

     

     

    2. Are there recommended best practices for stabilizing Launch Outlook in unattended runs?

    Yes. Microsoft and community best practices include (list is bit long):

    • Kill any existing Outlook processes before starting to avoid hung instances.

    • Start Outlook with the /recycle switch to reuse existing instances and reduce startup overhead.

    • Add explicit waits (5–10 seconds) after starting Outlook to allow full initialization.

    • Implement retry logic around the Launch Outlook action (e.g., 2–3 retries with delays).

    • Run PAD and Outlook under the same user context with proper permissions.

    • Avoid running flows when the VM session is locked or disconnected.

    • Disable unnecessary Outlook add-ins that may delay startup.

    • Keep Outlook and Office updated and repair installations if needed.

    • Use classic Outlook instead of New Outlook for better automation compatibility.

    Referecen is here

     

    3. Is there a known issue where Launch Outlook fails intermittently after the machine is idle / after nightly session recycling?

    Yes. It is a known pattern that after overnight idle periods or session recycling, the Outlook COM server may not initialize properly on first launch. The first run of the day or after a cold start often fails with COM exceptions due to profile loading, authentication, or COM registration delays.

    4. If you’ve solved this in enterprise environments, what changes made the biggest impact?

    Common impactful changes include:

    • Using /recycle switch when launching Outlook.

    • Implementing robust retry and wait logic around Launch Outlook.

    • Ensuring PAD runtime and Outlook run under the same user account with proper permissions.

    • Disabling unnecessary or problematic Outlook add-ins.

    • Configuring Outlook in Cached Exchange Mode to improve profile load times.

    • Keeping Outlook and Office fully patched and repaired.

    • Avoiding locked or disconnected VM sessions during unattended runs.

    • Using classic Outlook instead of New Outlook for better automation compatibility.

    • Creating warm-up flows that launch Outlook early in the day to avoid cold start issues.

     

     

    I am sure some clues I tried to give. If these clues help to resolve the issue brought you by here, please don't forget to check the box Does this answer your question? At the same time, I am pretty sure you have liked the response!
  • TM-18062219-0 Profile Picture
    13 on at
    Thank you very much @Haque for the response, I can see this is a Copilot/ChatGPT response that you've generated and I did try a lot of those things, and sadly they are not helping, so i was wondering if anyone else had come across this issue of launching Outlook unattended on a VM with Power Automate Desktop as there seems to be nothing online regarding this.
     
    What did come into my head trying would be to create a flow which manually clicks on the Outlook icon to launch it, to simulate how a user would do it, and maybe that would "warm-up" the VM into triggering correctly and allow it to work consistently for days on end. I did ask how UiPath and Blue Prism do this because I've used them before on VMs and never had an issue like this, and being a Microsoft product you'd think it would be much smoother to interact

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

Season of Sharing Community Challenge Launch!

Jump in, show your community spirit, and win prizes!

Kudos to our 2025 Community Spotlight Honorees

Expanding mentorship, skilling, and AI innovation

Congratulations to the May Top 10 Community Leaders!

These are the community rock stars!

Leaderboard > Power Automate

#1
Valantis Profile Picture

Valantis 377

#2
11manish Profile Picture

11manish 279

#3
David_MA Profile Picture

David_MA 234 Super User 2026 Season 1

Last 30 days Overall leaderboard