Perplexity Launches 'AI Computer' for Workflows
AI startup Perplexity has launched “Perplexity Computer,” a multi-model AI system that can autonomously generate and execute entire business workflows. The platform aims to create AI-powered digital workers that can handle complex processes like research and reporting without human intervention.
Perplexity Computer functions as an orchestrator, coordinating 19 different specialized AI models rather than relying on a single one. It assigns tasks to the best-suited model, using Anthropic's Opus 4.6 as the core reasoning engine, Google's Gemini for in-depth research, and others for tasks like image and video generation. The system is designed to execute complex workflows that can run for hours or even months without direct human input. It operates by breaking a user's stated goal into a series of subtasks, which are then carried out by autonomous sub-agents in a secure, isolated cloud environment equipped with a filesystem and browser. Founded in August 2022, the San Francisco-based company has seen rapid growth, raising more than $1.2 billion in funding and achieving a valuation of $20 billion. Its backers include high-profile investors such as Jeff Bezos, Nvidia, IVP, and SoftBank Vision Fund 2. The launch marks a strategic pivot for Perplexity, moving it from an "answer engine" competing with search giants to an "action engine" designed to provide deliverables and automate labor. The company's CEO, Aravind Srinivas, previously held research positions at OpenAI, Google Brain, and DeepMind before co-founding Perplexity. This new platform is being compared to other emerging autonomous AI agents like OpenClaw, with the potential to impact jobs involving repetitive digital tasks such as data entry and report generation. Access is initially available to Perplexity's premium "Max" subscribers for a monthly fee and operates on a credit-based system for executing tasks.
