The world of AI assistants is evolving from one-off question-and-answer bots to something far more integrated. Perplexity's new 'Teammates' feature is a major step in this direction, offering persistent AI agents that own ongoing work.
What Are Perplexity Teammates?
Think of Teammates
as specialized AI agents you can assign to a long-running job. Instead of giving a single prompt for an immediate task, you give a Teammate an entire responsibility, like managing a marketing campaign, tracking bug reports, or researching a new market. The Teammate then works autonomously in the background, planning tasks, bringing in other agents if the work expands, and seeing the project through until it's ready for your review. This shifts the model from a reactive assistant to a proactive, persistent collaborator that takes ownership of a workflow. Each Teammate operates from a dedicated, persistent thread in your workspace, making it easy to track its progress and review its work.
Building on a Foundation of Knowledge
The key to Teammates' power is its ability to build upon existing context. The feature is designed to integrate deeply with 'Perplexity Computer', the platform's engine for handling complex, multi-step tasks. This means a Teammate doesn't start from scratch. It inherits your existing projects, files, skills, and memory. For example, a sales-focused Teammate can use a project's stored account research, follow a pre-defined 'skill' for writing outreach emails, and run a weekly automation to review the sales pipeline. This ability to draw from shared documents, task histories, and user feedback allows it to maintain context over long periods, a crucial factor in reducing the friction often found in enterprise AI tools.
From Generalist to Specialist
With Teammates, Perplexity is pushing AI beyond a general-purpose tool into a collection of specialized roles. You can create a marketing Teammate to handle social media and SEO, a sales Teammate to find new opportunities in past customer conversations, or an operations Teammate to manage project deadlines. Each Teammate runs in its own isolated environment with its own browser and desktop access, ensuring security and separation between tasks. They can even be assigned different AI models—for instance, using a powerful model for complex planning and a faster, cheaper one for routine tasks—to balance cost and quality.
Proactive and Collaborative Workflows
A significant leap forward is the proactive nature of Teammates. They are designed to monitor connected tools and act when something changes without waiting for a command. A sales Teammate might notice a regular customer has stopped buying, investigate their order history, and then draft a personalized offer for your approval. This proactive monitoring extends to collaboration. Teammates can be added to Slack channels, where human colleagues can assign them work or ask for updates. Furthermore, different Teammates can collaborate, passing work to one another. One agent could identify bugs on a website and then hand off its findings to another Teammate tasked with preparing the fixes.
A Strategic Move in the AI Race
The launch of Teammates positions Perplexity directly against other major players like OpenAI and Anthropic in the competitive enterprise AI agent market. By focusing on persistent, autonomous agents that integrate deeply into a user's workflow, Perplexity is carving out a niche focused on longterm, goaloriented work rather than simple chat interactions. The system is designed to reduce the manual oversight needed for recurring business processes, allowing employees to focus more on strategy and final approvals. Initially available to subscribers of its hightier Perplexity Max plan, Teammates represents a vision for the future of work where every employee might have a team of AI agents working alongside them.
