Perplexity launches its own take on AI agents for work

Perplexity launched its own system of work-focused artificial intelligence agents called Teammates on November 5, 2026, entering direct competition with workspace automation offerings from OpenAI and Anthropic. Designed to operate around the clock without manual intervention, the always-on software agents take on distinct operational roles inside an organization while drawing directly from a user's existing projects, accumulated memory, and established skills across the Perplexity software environment. According to The Verge's report, the rollout positions Perplexity alongside OpenAI's Dots and Claude Duet by offering automated digital coworkers capable of handling complex workstreams twenty-four hours a day, seven days a week.

This breakdown examines the technical structure of the Teammates rollout, the financial requirements for access, and how the platform integrates with existing desktop automation infrastructure. The analysis is structured for enterprise leaders, engineering teams, and software operators evaluating multi-agent orchestration tools for office environments. By connecting existing workspace context with automated local execution, the release outlines how Perplexity plans to transition from a consumer-facing research engine into a persistent enterprise operational environment where automated agents execute concurrent assignments across corporate communication channels.

Perplexity its own take on AI agents for work: the announcement

The official release marks Perplexity's formal entry into persistent agentic workplace software, targeting operational workflows previously handled by human staff or single-turn prompts. The new Teammates tool operates on an always-on architecture, allowing individual agents to run 24/7 across multiple functional roles within an organization's digital workspace. Rather than requiring users to manually re-enter background data for every query, Teammates automatically connects to the Perplexity applications already deployed across a user's workflow, pulling historical projects, specialized skills, and accumulated memory directly into ongoing assignments without separate configuration.

By launching Teammates, Perplexity is directly challenging rival workspace systems, specifically OpenAI's Dots and Anthropic's Claude Duet, which similarly deploy autonomous agents to handle continuous organizational duties. Perplexity positions the product as an enterprise-grade collaboration layer where different automated roles run simultaneously across shared projects. To coordinate this multi-agent workload, Perplexity Chief Executive Officer Aravind Srinivas outlined an expansive corporate deployment vision on X, stating that over time, the company envisions every organization having hundreds of teammates working for each person within the enterprise.

What actually changed with Perplexity its own take on AI agents for work

The technical foundation of Teammates represents a significant architectural shift from single-action assistants to multi-threaded collaborative agents that build on top of Perplexity Computer. Perplexity Computer previously established the company's local automation footprint by deploying software agents directly onto user devices to carry out specific on-machine operations. Teammates expands that baseline capability by coordinating ongoing, large-scale work across distributed systems, enabling multiple agents to divide operational responsibilities and execute complex assignments in parallel rather than executing single commands sequentially.

Under this updated operational model, users can spin up separate Teammates for distinct business functions and assign each agent different jobs that execute simultaneously. Each Teammate maintains dedicated workspace context, drawing from saved files, project assets, and user instructions while operating autonomously in the background. Because the agents run continuously rather than timing out after finishing an individual task, they can monitor inputs, update files, and maintain long-running operational pipelines across the enterprise workspace without requiring persistent manual oversight from human staff.

Who should care about Perplexity its own take on AI agents for work

Corporate decision-makers, team managers, and enterprise operations professionals evaluating autonomous workplace tooling represent the primary audience for this release. Teams currently weighing agentic platforms must weigh Perplexity's unified workspace integration against competing solutions from OpenAI and Anthropic. Because Teammates ingests existing workspace history, teams that already store research files, technical documentation, and project briefs inside Perplexity gain immediate leverage, as new agents can immediately access prior context without undergoing manual onboarding or pipeline reconfiguration.

At the same time, budget-conscious organizations and independent developers must carefully review the commercial terms attached to the deployment. Perplexity has positioned the feature as a premium enterprise capability, pricing access considerably higher than comparable automation tools across the broader software industry. Teams managing high-volume, multi-threaded task loads across communication platforms will find the automated multi-agent structure attractive, but the ongoing recurring expense requires technical directors to verify measurable productivity gains before rolling the software out across departments.

How to try Perplexity its own take on AI agents for work

Access to Teammates requires an active subscription to Perplexity's top-tier Max plan, which costs $200 per month. The company has restricted the feature exclusively to this $200 monthly tier, establishing a substantially higher barrier to entry than standard consumer or business tiers across the generative artificial intelligence market. Prospective users must secure a Max tier seat before the software unlocks the multi-agent dashboard, project history routing, and the persistent background runtime environments necessary to maintain 24/7 agent execution.

Once subscribed to the Max tier, operators interact with their automated workforce through two primary interfaces. Within the native Perplexity application, each individual Teammate receives its own dedicated conversational thread, allowing users to inspect ongoing progress, issue modified instructions, and audit active assignments in isolation. For organizational coordination outside the standalone app, Perplexity integrates Teammates directly into Slack, enabling team members to ping agents, delegate new tasks, and review project milestones inside their existing team messaging channels.

What to watch after Perplexity its own take on AI agents for work

Industry attention will now focus on how effectively Perplexity executes its vision of dense, multi-agent workspaces where hundreds of virtual assistants assist individual corporate employees. Observers will track whether enterprise customers embrace the steep $200 monthly per-seat commitment, particularly when evaluated against OpenAI's Dots and Claude Duet in real-world deployment settings. The primary technical benchmark will center on how reliably Teammates handles long-running, large-scale assignments without drifting from original instructions or corrupting shared project memory across extended execution windows.

Further scrutiny will monitor the operational interplay between Teammates and the underlying Perplexity Computer device system. Because Perplexity Computer deploys automation directly onto physical hardware, administrators must assess how device-level execution, local resource consumption, and centralized Slack collaboration interact during continuous 24/7 operations. As organizations deploy multiple agents simultaneously, monitoring execution reliability, network stability, and coordination across team messaging channels will determine whether Perplexity's premium agent tier succeeds across enterprise office environments.