Gone are the days when you had to write code to automate your browser using tools like Playwright. Microsoft has now completely changed the game with its new tool called Scout. It is an always-on AI agent that Microsoft introduced for Microsoft 365 at Build 2026.
Using this tool, you can now describe your task in plain language instead of writing code, and it will drive a real Playwright-controlled browser on your behalf. However, if you aren’t sure how to get started with Microsoft Scout, then this article will explain it all below.
What is headless web scraping?
Headless web scraping is the process of extracting data from websites using a web browser that runs without a graphical user interface (GUI). This is where, Microsoft’s Playwright can be used for controlling Chromium, Firefox, and WebKit in headless mode, allowing you to automate browsing, scrape data, and run tests efficiently at scale.
With Microsoft’s Scout and its Playwright integration through the Model Context Protocol (MCP), you can describe tasks in natural language instead of writing browser automation code. Scout then controls a real Playwright-powered browser on your behalf, making web automation and scraping more accessible.
How to Web Scrape with Microsoft Scout and Playwright
Before you start controlling your browser or scraping the web, you need to enroll and enable a few things since browser control isn’t turned on for every user by default. To get it enabled, you will need to:
- Make sure to enroll in Microsoft’s Frontier Preview Program.
- Enable it through the Microsoft Admin Center under Agent Management.
- Turn on Browser Control: Open Scout > Settings > Permissions > Toggle on Browser Control and Web Browsing.
Once you are done with the above steps, you can get started with Scout. As there is no need to write code, you can simply ask in plain language, such as:
- Go to the Azure portal and check my resource group’s cost this month.
- Fill out the contact form with this information:
- Take a screenshot of the pricing page.
- Extract the table on this page and turn it into a CSV.
So that was a quick guide on how to get started with browser control using Scout. If you have used any AI assistant tool before and know anything about prompting, you shouldn’t really face any challenges using the tool. Just make sure you are writing good prompts.
How Microsoft Scout + Playwright MCP Works
MCP is a common interface that helps AI models like Copilot, GPT-4, and Claude control Playwright without writing custom integration code. In this case, the snapshot mode uses the accessibility tree to read website content and the screenshot mode is useful for tasks where the AI needs to understand the visual layout of the website.
Limitations to keep in mind
- Scout remains a Frontier preview feature.
- Access depends on organisational and administrator settings.
- Administrators can block Playwright, restrict permissions or prevent Scout from accessing particular websites.
- Login challenges, CAPTCHAs and frequently changing page layouts can affect browser automation.
- Users should review extracted data before relying on it.
Microsoft allows administrators to block browser destinations, disable Playwright and require approval for actions.
