Perplexity Launches Workflow-Executing AI

Perplexity Computer functions as a "digital worker" by coordinating dozens of specialized AI models to operate software and execute multi-step workflows just as a human would. Instead of a user clicking through different tabs and tools, they describe the desired outcome, and the system breaks it down into tasks, assigning them to the most appropriate AI models for execution. This allows it to handle complex processes that can run for hours or even months. The system uses an orchestration layer to manage various leading AI models simultaneously. For instance, it might use Anthropic's Opus 4.6 for core reasoning, Gemini for in-depth research, and ChatGPT 5.2 for tasks requiring long-context recall. This multi-model approach is based on the idea that specialized models are more effective for specific tasks than a single, universal model. The development of Perplexity Computer signals a strategic shift for the company, moving beyond its reputation as an "answer engine" that provides cited search results. This evolution into a system that performs actions reflects a broader industry trend where AI is moving from an advisory role to an operational one. The new platform is initially available to subscribers of Perplexity's "Max" tier. Founded in 2022, Perplexity was established by engineers with backgrounds at prominent AI labs like OpenAI, Google Brain, and DeepMind. CEO Aravind Srinivas, who previously worked at all three, co-founded the company with the initial goal of redesigning online search for the AI era. The company has seen rapid financial growth, with its valuation jumping significantly in a short period. After a series of funding rounds in 2024 and 2025, the company's valuation reached an estimated $20 billion by September 2025. Key investors include Nvidia, Jeff Bezos, and SoftBank Vision Fund 2.