UI Flows made their public preview last week, which seemed the ideal occasion to play around with this new and promising flow of the Microsoft Power Platform. Not familiar with the concept of UI flows? Check out this video, which explains perfectly what an UI Flow is capable of and which use cases are applicable. Sometimes they are also referred to as RPA (robotic processing automation).


Where other flows are plain and simple to start with (you only need a browser), UI Flows require some additional components to be installed. However, Microsoft makes this very easy. You need to install the Microsoft Flow Extension to the browser. Only 2 browsers are supported at time of writing: Google Chrome and Edge (Chromium). The last one is still in beta at time of writing, so I used the Chrome browser for my tests.

If you want to use a web app, you will also need to install the selenium IDE extension to your browser. And finally, you will also need to have a on-premise data gateway installed and configured to be able to access your UI Flow from another flow (more on that later).

Building UI Flows

Installing and configuring the browser extensions went really smooth. The only thing I had a real pain with, is to get it started. Since I am used to work into the “make.powerapps.com” portal and navigate to flows over there, the Flow extension wasn’t triggered. So tip one is to start and build a UI flow from “flow.microsoft.com”, which might seem obvious to a lot of you, but it wasn’t for me.

Once you are able to start the recorder, you will see a screen capture like UI on top of your screen with some buttons to insert input fields or to capture output fields. I found this very easy to use for the desktop flow. It highlights fields with a blue border and you can then choose to either capture or insert values from your field. When you’re done recording, your script is automatically saved to your flow.

For web apps on the other hand, things are a bit more complicated. When you launch the recorder, it will open op the Selenium extension. People who are not familiar with this IDE will have some struggle to find out how to use it. You press the record button and the base url will be opened (which you provided when creating the web app). Now Selenium will capture all your actions. To add input fields, you will need to add commands to the selenium test script yourself. The value is passed in following format: ${value}. This works as it should and when consuming this UI Flow, the input parameter was automatically added.

Bringing it together

As previously mentioned, UI flows are addressed via a gateway. When you want to add an UI Flow, Power Automate automatically helps you by providing the option to add a new on-premise gateway. Installation is straightforward and you will need to provide connection credentials which can be reused in your flows to access your on-premise applications.


In the end, I really like this addition to Power Automate. It doesn’t require any changes to legacy or web applications used by organizations, which makes the step to the Power Platform a bit easier. Just be aware RPA flows require some software being installed and you should familiarize yourself with Selenium. Also, actually one browser is supported at the moment, which is Google Chrome (the edge browser is still in beta).

Microsoft did a very good job on the documentation too. It’s clear explained with some nice examples how to get you started an bringing all the dots together.