Perplexity Unveils AI Teammates for Professional Workflows

How Teammates Operate Within Existing Workspaces

San Francisco-based AI startup Perplexity has released a new suite of autonomous agents designed specifically for business environments. The product, branded as Teammates, functions as persistent digital workers that can assume distinct roles within a corporate workspace. These agents integrate directly with existing Perplexity tools, inheriting prior project data, learned skills, and contextual memory.

The launch positions Perplexity in direct competition with OpenAI and Anthropic, both of whom have recently introduced their own enterprise agent products. Perplexity's approach emphasizes continuity, allowing the agents to build on knowledge already stored in the company's ecosystem rather than starting from scratch.

Unlike standalone chatbots, Teammates are designed to remain active across multiple sessions. They can execute multi-step tasks, coordinate with human colleagues, and adapt their behavior based on accumulated memory. The system pulls from the user's existing Perplexity projects, meaning an agent can reference past work without requiring manual context input. This reduces setup friction for teams already using the platform.

Why This Matters in the Enterprise AI Race

The agents are role-specific, meaning a company can deploy a researcher, a coder, or a project manager agent, each with tailored capabilities. Perplexity states the agents can collaborate with each other and with human team members, functioning as true digital coworkers rather than simple command-and-response tools.

The enterprise AI agent market is heating up rapidly. OpenAI and Anthropic have both rolled out competing products targeting the same professional use cases. Perplexity's differentiator is its memory and integration layer, which leverages the company's existing search and knowledge tools.

Analysts note that the key battleground will be reliability and trust. Companies need agents that can be left running unattended, which requires robust error handling and audit trails. Perplexity's emphasis on persistent memory addresses a common pain point: agents that forget context between sessions.

Frequently Asked Questions

What makes Perplexity's Teammates different from OpenAI or Anthropic agents? Perplexity's agents inherit memory and skills from existing Perplexity projects, providing continuity that competitors' products may lack. They are also designed to operate as always-on digital coworkers within a workspace.

Can Teammates work independently without human oversight? Yes. The agents are built to handle multi-step workflows autonomously, though Perplexity recommends human review for critical decisions. They can coordinate with each other and with human team members.

Is this product available now? Perplexity has made the Teammates product available to its enterprise customers. The release targets organizations already using Perplexity's search and knowledge tools, with integration built in from the start.