Lesson 187 · 2 min
Browser agents: AI that uses websites for you
A browser agent looks at web pages, decides where to click and type, and repeats until a task is done, a handy trick that is slower and riskier than it looks.
Many websites have no connector or API. A browser agent works around that by using the site like a person. It reads the page, sometimes from a screenshot, decides on an action such as "click Search" or "type the date", does it, and looks again.
That loop lets it handle tasks like checking prices on several shops, filling a long online form or collecting details from many listings into a table.
It is also slow and fragile. Pages change, pop-ups get in the way and buttons move. An agent can misread a page and click the wrong thing, and every page it reads could contain text written to mislead it.
Use browser agents for low-risk, repetitive web chores you can check. Run them in a separate browser profile without your important logins when possible, and require a pause before any payment, booking, message or deletion.
Do it now
Plan a browser chore
- 1. Think of a boring web task you repeat, like checking three sites for a price.
- 2. Write the steps a person would take, click by click.
- 3. Mark the step where a mistake would matter most, and add "stop and ask me" there.
I want a browser agent to [TASK]. Break it into click-level steps. Mark any step that could cost money, share data or be hard to undo, and add a check with me before each of those.
Quick check
Words from this lesson
- Agent
- An AI system that takes several steps on its own, such as searching, clicking or using apps, to finish a goal you set.
- Computer vision
- AI that makes sense of images and video, such as reading signs, spotting faces or checking products on a factory line.
- Prompt injection
- Hidden instructions planted in a web page, email or file that try to hijack an AI reading it.
- Human in the loop
- Designing a process so a person reviews or approves the AI's work before it has real consequences.