Perplexity's browser strategy
Perplexity at $148M/year
Owning the browser matters because it lets Perplexity move from answering a question after intent is expressed to capturing intent at the exact moment a user starts doing work. Inside a browser, the product can see the page, pull context from tabs and forms, and turn a research session into actions like booking, buying, or submitting, which is a much more defensible position than being one more search box with citations.
-
Browser control is a distribution and data play. Research on AI agent form factors shows that companies move lower in the stack when they want more context and more ability to act. A browser sits closer to the user than a search page does, so it sees what site the user is on and can act without waiting for a separate query.
-
The practical product difference is simple. A search app returns an answer. An agentic browser can read the current page, click buttons, fill forms, upload files, and carry a task across multiple sites. That is why browser automation is becoming the bridge between AI reasoning and real world workflows, especially where normal APIs do not exist.
-
The competitive set also changes. Perplexity is no longer just compared with answer engines, it is competing for the interface layer against OpenAI and Google products that also combine browsing with action taking. That makes retention less about who has a slightly better model, and more about who becomes the default place people start and finish tasks on the web.
From here, browsers become the main wedge for consumer AI agents because they combine intent, context, and execution in one place. If Perplexity can make Comet the default environment for research, shopping, travel, and other repeat workflows, it can turn occasional search usage into daily task completion and build a much stickier product than search alone.