Perplexity Launches Workflow AI
Perplexity launched "Computer," a multi-model AI system that creates and executes entire workflows autonomously. The tool aims to automate complex, multi-step tasks for founders and teams. A user reported that automating Google Keep and Tasks with Gemini has been a "game-changer," saving significant time through auto-scheduling and tagging.
Perplexity's "Computer" operates as a multi-model orchestration system, breaking down complex user requests into smaller tasks and assigning them to the most suitable specialized AI models for execution. This allows it to tackle multi-step projects like market analysis or generating detailed reports by leveraging the unique strengths of various AI systems in a coordinated fashion. The system is designed to be model-agnostic, meaning it can integrate and choose from a variety of leading AI models. For instance, it might use Google's Gemini for in-depth research, Anthropic's Claude for tasks requiring a large context window, and OpenAI's GPT models for creative text generation, all within a single workflow. This venture is led by CEO Aravind Srinivas, who previously conducted AI research at Google, OpenAI, and DeepMind. The founding team also includes CTO Denis Yarats, a former AI research scientist at Meta, and Chief Strategy Officer Johnny Ho, previously an engineer at Quora. Founded in 2022, the San Francisco-based startup has seen a meteoric rise in valuation. After multiple funding rounds, Perplexity's valuation has surged, with significant investments from figures like Jeff Bezos and major tech companies including Nvidia and Databricks. The launch of "Computer" places Perplexity in direct competition with other AI-driven productivity tools and platforms from giants like Google (Gemini) and OpenAI (ChatGPT), as well as other notable players like Anthropic (Claude). Perplexity's core mission extends beyond simple conversational answers, aiming to create an "answer engine" that provides direct, synthesized information. The introduction of autonomous workflow execution marks a significant step towards a more agentic and capable AI system that can manage complex tasks over extended periods.
