A Browser That Buys You Flowers? Opera Just Changed the Web Forever

In March 2025, at the Mobile World Congress in Barcelona, Opera unveiled what many are calling the most significant leap in web technology since the rise of the browser itself: the world’s first truly agentic browser. Dubbed the Opera Browser Operator, this AI-driven platform doesn’t just assist users — it acts on their behalf, autonomously navigating the web, interpreting complex tasks, and completing real-world actions with minimal input. The moment that sealed its legacy came in April during a live demonstration in Lisbon. With a simple command — “Buy me flowers for my mother’s birthday” — the browser instantly began executing the task. It searched for florists in Lisbon, filtered results by reputation and delivery time, selected a local shop, translated the site’s Portuguese content, filled out the order form, processed payment through a secure portal, and confirmed the purchase — all within minutes, without any further human intervention. This wasn’t a pre-recorded stunt. It was real-time, end-to-end automation, powered by deep, low-level access to the web’s Document Object Model (DOM), not just screenshots or APIs. This architectural choice gives Opera’s AI a critical edge: it can interact with websites exactly as a human would, reading dynamic content, clicking buttons, and handling forms with precision — far faster and more reliably than systems relying on image-based AI interpretation. Compared to competitors, Opera’s approach stands out. Perplexity’s AI assistant still functions primarily as a search tool with limited automation. OpenAI’s browser extensions remain task-specific and require frequent user oversight. Google, while investing heavily in AI-driven browsing, has yet to launch a fully autonomous agent. Opera’s Browser Operator, by contrast, operates as a persistent, goal-driven agent that can manage multi-step workflows across domains. Yet this leap forward brings profound new challenges. With the ability to act on behalf of users, agentic browsers open a vast new attack surface. If compromised, an AI agent could make unauthorized purchases, access sensitive accounts, or manipulate data without detection. Unlike traditional malware, such threats could operate silently, mimicking legitimate user behavior. Privacy concerns are equally pressing. The browser’s deep access to websites means it can collect and process vast amounts of personal data — from login credentials to financial details — even if users believe they’re only giving a single command. While Opera claims to prioritize privacy with on-device processing and end-to-end encryption, the model’s reliance on real-time web interaction makes full transparency difficult. More than a tool, the Opera Browser Operator signals the arrival of Web 4.0 — the Agentic Web. Here, the internet is no longer a passive information space but a dynamic ecosystem where AI agents perform tasks, negotiate services, and manage digital lives. Human interaction shifts from direct control to oversight and intent-setting, fundamentally altering how we engage with technology. This isn’t just a browser upgrade. It’s the beginning of a new era — one where the web doesn’t just respond to us, but acts for us. And as with every great leap in technology, the question isn’t just what it can do, but who controls it, and what happens when it goes wrong.