Perplexity just turned search into a build-while-you-sleep machine.
The Summary
- Perplexity announced Perplexity Computer, a new product that moves beyond answering questions to executing multi-step tasks autonomously
- The shift represents AI search evolving from information retrieval to agent-driven action, competing directly with Google's search monopoly
- CBO Dmitry Shevelenko positioned this as search infrastructure for the agent economy, not just better search results
The Signal
Perplexity isn't playing the search game anymore. They're building the operating system for AI agents that actually do things. The Perplexity Computer announcement signals a fundamental product pivot from "answer engine" to "action engine." This is the search company betting that asking questions is just the warm-up act.
The product architecture matters here. Traditional search gives you ten blue links. Perplexity's first iteration gave you synthesized answers with citations. Perplexity Computer goes three steps further: it understands intent, plans multi-step workflows, and executes them. Book the flight, file the expense report, schedule the follow-up. The question becomes the job description.
"Search is evolving from information retrieval to agent-driven action."
This puts Perplexity in direct competition with every productivity tool, not just Google. If your agent can research, plan, and execute in one interface, why do you need separate tools for project management, data analysis, or customer research? The answer-engine-to-agent pipeline is the same path OpenAI is running with ChatGPT, Anthropic with Claude, and Google with Gemini.
The timing connects to something bigger. We're six months past the agent hype cycle peak, and the companies still standing are the ones that picked specific jobs to automate. Perplexity Computer is betting on search-as-starting-point for agentic workflows. That's smart positioning. Every knowledge work task starts with a question. If Perplexity owns the question-to-action pipeline, they own the front door to the agent economy.
Key competitive dynamics:
- Google has distribution but legacy search infrastructure
- OpenAI has model capability but no search muscle memory
- Perplexity has search UX and the ability to ship fast
The business model implications run deep. Perplexity's revenue today comes from subscriptions to better search. Perplexity Computer subscriptions will be priced on actions completed, not questions answered. That's a different value proposition. It's also a different competitive moat. The best action engine won't be the one with the best model. It'll be the one with the deepest integration catalog and the most reliable execution layer.
The Implication
Watch how Perplexity prices this. If it's per-task execution, they're building usage-based SaaS for the agent economy. If it's flat subscription with unlimited actions, they're betting on becoming essential infrastructure. Either way, the race is on to own the layer between "I need to do this" and "it's done."
For knowledge workers, the question is simple: what parts of your job start with research and end with a repeatable workflow? That's what gets automated next. Start mapping those workflows now, because the tools to replace them are shipping.