Perplexity Launches Teammates: AI Agents for Work

Perplexity has entered the enterprise AI agent market with Teammates, a new system of always-on agents designed to handle workplace tasks across multiple roles. Officially launched on November 5, 2026, Teammates positions the startup as a direct challenger to OpenAI and Anthropic in the rapidly expanding space of work-focused artificial intelligence. The system integrates with Perplexity Computer, allowing agents to operate directly on user devices while accessing existing projects, skills, and organizational memory.

What Perplexity Teammates Offers

Teammates delivers a multi-agent workspace where individual AI agents can assume different organizational roles simultaneously. Unlike single-purpose assistants, each Teammate can handle specialized functions within a broader workflow, from research and analysis to project management and communication tasks. The system pulls context from tools users already employ within the Perplexity ecosystem, maintaining continuity across tasks and reducing the need for manual context-setting.

The integration with Perplexity Computer represents a technical shift in how these agents operate. Rather than functioning solely through cloud-based APIs, Teammates can deploy directly onto user devices, enabling faster response times and greater control over data handling. This architecture gives organizations more flexibility in how they manage sensitive information while maintaining the capabilities of cloud-connected AI systems.

Competitive Positioning Against OpenAI and Anthropic

The launch arrives as OpenAI and Anthropic have both expanded their enterprise offerings with agent-based systems. Perplexity's approach emphasizes persistent, role-based agents that remain active within a workspace rather than requiring per-task invocation. This design choice targets organizations seeking continuous AI support rather than episodic assistance.

Key differentiators include:

• Always-on availability across workspace contexts
• Native integration with Perplexity's existing search and knowledge infrastructure
• Device-level deployment through Perplexity Computer for enhanced privacy controls
• Multi-role capability within a single workspace environment

Technical Architecture and Integration

Teammates leverages Perplexity's core search and retrieval technology to ground agent responses in current information. The system maintains organizational memory across sessions, allowing agents to reference past projects, decisions, and workflows without manual prompting. This persistence addresses a common friction point in enterprise AI adoption, where context loss between interactions reduces practical utility.

The Perplexity Computer integration enables agents to access local files, applications, and system resources when appropriate, extending their capabilities beyond text-based interactions. Organizations can configure permission levels and data access rules to align with security policies and compliance requirements.

Release Date and Availability

Officially launched on November 5, 2026, Teammates is now available to Perplexity enterprise customers. The company has not disclosed detailed pricing structures or tiered access levels publicly, though the system appears positioned as a premium offering within Perplexity's product suite. Organizations interested in deployment can access Teammates through Perplexity's enterprise portal, with configuration options for workspace size and agent role distribution.

What This Means

Perplexity's entry into work-focused AI agents signals growing competition in enterprise artificial intelligence beyond conversational interfaces. By combining always-on availability, role-based specialization, and deep integration with its existing infrastructure, Perplexity offers organizations an alternative to incumbent solutions. The success of Teammates will likely depend on how effectively it reduces friction in real-world workflows and whether its persistent-agent model proves more valuable than on-demand assistance. As enterprises continue evaluating AI deployment strategies, product differentiation in this space will increasingly hinge on practical integration depth rather than raw capability benchmarks.