Powerful Devs Conference Questions

[Shavonna]

Hi all! Thank you so much for attending the Powerful Devs Conference 2025. Please use this space to ask any remaining questions you may have pertaining to the conference.

If you enjoyed the conference or have any feedback, please fill out our Powerful Devs Conference Survey!

If you missed the conference, you can watch it as well as the Hack Together Livestreams here.

[mitarana25]

Hi I am interested in in depth security roles demo in related to copilot and model driven apps.

[ziptechBowwi]

The earlier demo showcased PCF and GitHub Copilot—can Copilot assist in debugging YAML?

[aprildunnam]

Good question - yes I use it all the time to debug YAML!
https://github.com/user-attachments/assets/ef7500db-53ec-4180-a211-a84f245dcabc

[PRusakPL]

Great presentation from Lauren Mills (full of knowledge) 💪🌞

Question about how actions that require slot filling will be charged (Lauren showed it can be done in 2 ways):

• Classic Orchestration - usage of Plugin Action inside topic
• Generative Orchestration - this will be magically working by itself

Will all actions that require slot filling cost 25 messages (treated as Autonomous Actions) ?

[lauren-mills]

Actions called from a Topic when in Classic mode are not currently charged as Autonomous Actions, but as regular messages.

[lauren-mills]

However, there may be changes here in the future.

[sskninetyfour]

​​what's the best way to test JavaScript on ribbon/command bar buttons?

[scottdurow]

Hi @sskninetyfour - thank you for posting this question!

In broad terms, there are 3 techniques:

1. Use the command checker - https://learn.microsoft.com/en-us/troubleshoot/power-platform/power-apps/create-and-use-apps/ribbon-issues
2. Use F12 Developer tools in Edge/Chrome https://learn.microsoft.com/en-us/microsoft-edge/devtools-guide-chromium/javascript/
3. Write Unit Tests using a test framework such as Jest and a mocking framework

Is there a specific issue that you are having that I might be able to help with?

[Grant-Archibald-MS]

@sskninetyfour is the question related to testing the command bar of a Model Driven Application?

If so that is an example we are working for Test Engine which has the ability to test Model Driven Applications (MDA)

[scottdurow]

Thank you @Grant-Archibald-MS - I didn't clock that the question might be about automated testing! Ofc! 🙃

[DanielKerridge]

Really great live stream today - I am looking to get a copy of the 'scoring script' where you run the pipeline and then it gives you warnings if you have used the wrong naming convention in your app - could you point me towards that? (2hr 41min into the live stream) - I had a look on the github but could not find it - Cheers!

[appieschot]

Will wrap up asap to get it to you! Sorry for the delay!

[appieschot]

There you go, the full script is available from: https://www.cloudappie.nl/powerplatform-source-code-naming-tips/ let me know if you have any questions or can't get it working! Sorry again for the delay!

[strwase]

I unfortunately missed the conf yesterday. Is there any way for me to rewatch it?

[AlmasMahfooz]

https://www.powerfuldevs.com/Watch/

[nova-ac]

Can you provide details on built-in pipelines for solutions in Power Platform, specifically regarding their support for different deployment configurations for different environments similar to how the Center of Excellence (CoE) Application Lifecycle Management (ALM) supports different deployment configurations via the JSON config files?

[Grant-Archibald-MS]

@nova-da I will be adding examples of how I did the ALM examples I did for my session on Automated Testing (https://www.youtube.com/watch?v=vaq8UDUqkYw&list=PLlrxD0HtieHgu6fQ7watAyBglEZ8-oxg9&index=9) with tests and adding those examples to the Test Engine GitHub repository.

The basic process I showed at a high level:

• Development, Test and Production Environment configured with Pipeline for Power Platform in an environment
• A "Deployment" solution used to extend the pipeline to execute the Test Engine tests as part of an Azure DevOps Pipeline
• The Weather sample Model driven application that was being tested

[nova-ac]

@Grant-Archibald-MS Thanks for your reply! Just to confirm, you're referring to this repo. When do you think your examples will be added?

[Grant-Archibald-MS]

@nova-da yes that is the correct repo. We are working on series of release updates at the moment which will include a number of changes into the main branch.

The samples take advantage of the existing Extend pipelines in Power Platform to call to an Azure DevOps pipeline. However this is just a Cloud flow using the Azure DevOps connector. Other connectors to run in a GitHub runner or a another build agent could be considered as pac test run runs cross platform.

Once we have the new version of the pac test run command available then the Azure DevOps pipleline will be simpler. Over time we may look at two different versions of the examples that follow the Ring deployment model for test Engine

• Inner Ring This would need to build from source and require the .Net 8.0 SDK to be installed. It would have access to the most features including Experimental features but would be based on the Open Source Code
• Secondary Ring This will make use of pac test run to run the tests. In the initial phases it will require configuration in the test Settings to make use of Experimental features
• Outer Ring Will make use of pac test run with features that have migrated from the Experimental phase

[Grant-Archibald-MS]

I will update here as soon as the feature branch with the examples is available

[Grant-Archibald-MS]

As we look at the topic deeper we are also planning for the following:

• Getting started video - Show how to setup and author your tests in Windows or Linux
• Pipelines for Power Platform samples - Demonstrate how to deploy and extend pipelines with integration with Azure DevOps
• Learning Path Updates - Update to the hands on learning path to apply the concepts
• Security - Look deeper around security. How handle authentication across multiple machines from authoring and into CI/CD scenarios. Look at how factors like MFA and Conditional Access policy will affect your test setup and execution
• Providers - Look at how different providers beyond Power Apps fit into the Test Engine Model
• Extensibility - How to extend the Test Engine Providers or Power Fx functions to add new features or integrate with your existing testing investments
• Architecture - How this fits into Power Well Architected and how we extend Architecture Guidance with ALM examples

[wilmandx]

Hello, using power platform pipelines or even azure devops pipelines during presentations they recomends to use a service principal instead service accounts, but I wonder in large organisations with severals environnents is it not dangereaus that some power automates reach limits and affect other automates ? I was reading about limits and I saw for service principals there is a limit of 25.000 requests by 24h and requests can increase fast because each loop counts, variable initilization, etc. I add the Microsoft documentation and I would to understant better this limits for tenant request. https://learn.microsoft.com/en-us/power-platform/admin/api-request-limits-allocations#non-licensed-user-request-limits

[scottdurow]

Hi @wilmandx - thank you for posting this question and it's a really good one!

Yes, Service Principals are non-licensed users and draw from the base tenant request allocation. For cloud flows that are consuming a large number of requests, organizations will typically use PAYG, Associate the flow to an app, or use a Power Automate Process License

More information https://learn.microsoft.com/en-us/power-automate/change-cloud-flow-owner#change-the-owner-of-a-cloud-flow-to-a-service-principal-application-user

I hope this helps!

[shrutika-patil20]

Hello everyone! 👋

I have two questions, so instead of creating multiple posts, I’m combining them here:

1️⃣ First off, huge thanks to @gomomohapi for the amazing AI Builder demo! 🙌 I wanted to check—can AI Builder handle large prompts exceeding 16,384 tokens (in case using 4o-mini), or do I need to break them into small chunks?

2️⃣ When integrating AI capabilities into a Power Automate flow, which option is more powerful in terms of pricing, reliability, and overall performance—AI Builder or Generative Actions?

[gomomohapi]

Hi @shrutika-patil20

Firstly, thank you so much! 😊 I'm glad you liked the demo.

Now for your questions -

1️⃣ You'll see here that both models in AI Prompt Builder have a context window of 128k tokens. Context window includes both prompt and the model's response. So technically, the short answer to your question is yes. But if you're considering things like implementation, model response size, and performance - it's always best to break up your prompts for better processing if possible.

2️⃣ This is purely based on context as these two features do different things. Maybe if you can provide some additional notes on what you're trying to achieve?

[shrutika-patil20]

Thanks, @gomomohapi, for taking the time to respond to the query! 😊

For my second query—I’ve used AI Builder in a few projects, referencing it in Copilot Topics as an action. Right now, I’m experimenting with Autonomous Agents and exploring whether I can generate an answer directly within the Autonomous Agent flow by calling a Generative Action instead.

Maybe my approach is off, but I’d love to hear your thoughts!

[shrutika-patil20]

Hey everyone! 🎉👋

Wow! I just went through all 41 submitted projects, and I wanted to say—what an incredible display of creativity and learning! 🚀💡

Thank you Powerful Devs team for organizing such an amazing event. 🙌 This platform truly allows us to explore preview features, bring our ideas to life, and share them with the community.

Super excited to know when the results will be announced! 🏆✨

Thanks,
Shrutika Patil