How to use AI browsers at work
AI browsers are finally here. 5 ways to use them including: Market analysis in minutes, write documentation, teach yourself anything, keep up to date with competitors. Knowledge Series #76
🔒The Knowledge Series is available for paid subscribers. Get full ongoing access to 75+ explainers and AI tutorials to grow your technical knowledge at work. New guides added every month.
Last week, I installed the beta version of The Browser Company’s new product, Dia. As someone who never really managed to click with their previous product, Arc (despite trying very hard), I was a little sceptical.
Now, just a few days later I’m pretty much completely sold on the concept of AI browsers and find myself using the AI features far more than I had imagined. Using an AI browser has transformed how I interact with many of the pages I visit. When you give your browser “contextual awareness” of your tabs, documents and intentions, it suddenly opens a whole bunch of new ways to use a browser that I hadn’t really considered before trying it.
As well as Dia, Perplexity is working on its own upcoming AI browser, Chrome is gradually adding more AI features and Opera launched an AI browser of its own too.
In this Knowledge Series, we’ll explore how product teams can use AI browsers at work to transform common tasks and workflows like conducting market research, getting to grips with technical documentation, staying ahead of competitors and writing up PRDs / other documentation. We’ll also explore the brand new concept of AI browser “skills” that are gaining popularity right now and how you can use AI browsers to analyse data and spreadsheets in ways that Google Gemini and Copilot simply can’t match.
Coming up:
What is an AI browser? A simple overview of what they are and how they work
5 real world use cases for product teams including:
Conducting and automating market research
Writing PRDs, status reports and emails
Interacting with and understanding technical documentation
Teaching yourself about any new concept by talking with YouTube videos
Analyzing data in spreadsheets and multiple data sources
A comparison of all the AI browsers available right now with links to more info
What is an AI browser?
Before we look at some of the ways you can use these browsers, let’s first of all get a snapshot of some of the core technologies used in AI browsers and why they have the potential to transform how we use the web.
Here’s a brief overview of some of the main technologies involved in AI browsers:
Contextual awareness is at the heart of the power of AI browsers. A few days ago, Perplexity’s CEO made reference to this in relation to their upcoming new browser, Comet, with a simple tweet that read “context is all you need”:
In AI browsers, contextual awareness means using the information that the browser has access to in smart ways that allow users to transform every day workflows. For example, this means you can ask AI browsers to reference multiple tabs at the same time when performing competitor analysis (more on that later) or reference a set of brand guidelines when crafting a release note or email about a new upcoming feature.
Contextually aware browsers use the DOM and APIs to read web pages and combine this with LLMs to achieve pretty powerful results. If you want a recap on front end technologies and APIs you can read about these in other Knowledge Series here and here. They can “see” the pages you view and - with permission - use locally stored browsing histories to personalize your experience.
On-device machine learning helps the browser learn from user behaviors and fine tune the user experience by suggesting relevant websites based on previous behavior and AI models from OpenAI, Gemini, Meta etc are the engine of the browser’s smart capabilities, providing assistance throughout the entire process, typically through a sidepanel or chat interface.
Dia’s CEO says that its AI models will become more personalized over time:
With every tab that you open, it should feel like this AI model is getting more and more personalized to you, such that at the end of a week of browsing, a month of browsing, let alone a year, it’s going to know you as well as your closest friends and colleagues.
5 practical ways you can use AI browsers at work
Now let’s dig into some of the practical ways you can use AI browsers at work. For these examples, I’ve used Dia since this is one of the most fully fledged offerings (and Perplexity hasn’t released theirs just yet).
1.Conduct market research and competitor analysis
Contextual awareness means you can approach market research and competitor analysis in a slightly different way.
In this example, let’s imagine we’re working at a customer support SaaS company that is a direct competitor to tools like Intercom’s Fin or Decagon.
With these two browser tabs open, you can ask an AI browser to reference both of these open tabs and create a comparison table outlining the core features of the two products.
If you had a third tab with your own product open, you could then ask your AI assistant to identify feature gaps and potential opportunities for differentiation across each of these three products. Being able to actively use tabs in this way is pretty powerful.
Referencing other document types for customer research or accessibility checks
As well as tabs, you can also reference videos, documents or spreadsheets. We’ll explore spreadsheet analysis in later examples in a bit more detail but let’s imagine you have a document open which outlines your product roadmap in one tab and some customer research open in another. Your customer research could be in the form of a video hosted on YouTube or a document. Since your AI browser has access to both of these, you could ask it to check your product roadmap against insights from product research.
Accessibility checks are an opportunity here, too. In this example, I’ve visited a website that outlines accessibility checks and asked the browser to check an uploaded UI (in this case a snippet from Apple’s Liquid Glass on Apple Music) to see if it passes the criteria outlined.
Using custom skills as shortcuts for market analysis
Dia also comes with the ability to create custom skills as shortcuts. Skills allow you to quickly run prompts that you pre-program and run whenever you need them.
In this example, I create a prompt that asks the browser to search Reddit for threads on a specific topic relevant to my product and find problems, solutions and upvoted comments.
With this skill created, I can then invoke it whenever I want for any industry. In this example you can see how I can use this for industries like “pet insurance” or “saas customer service” products. Within a few seconds, you’ll get an up to date view of what users are saying on Reddit about your industry, just by using your keyword in the browser: