Perplexity has introduced Teammates, a new set of work-focused AI agents that the company says can stay active inside a workspace and handle different roles at the same time. Based on the information supplied in the source, the launch puts Perplexity more directly into competition with workplace agent efforts from OpenAI and Anthropic.
The Verge reports that Teammates are designed to pull from existing projects, skills, and memory across Perplexity tools already in use. Perplexity also says the product integrates with Perplexity Computer, its system for deploying agents on a user device to complete tasks, but positions Teammates as a layer for more collaborative and ongoing work.
According to the source, users can assign different jobs to different Teammates and have them work simultaneously. That suggests Perplexity is aiming beyond one-off prompts and toward multi-agent coordination for recurring office workflows.
What is confirmed from the supplied report is the product framing: always-on agents, multiple roles in one workspace, and links to existing Perplexity memory and projects. What remains unclear from the available details is how much autonomy Teammates have in practice, what controls managers get, and which enterprise systems are supported beyond Perplexity’s own tools.
The launch reflects a broader shift in AI products for work: vendors are trying to move from chat assistants toward delegated task execution. In practical terms, that could mean an AI system not just answering questions, but tracking a stream of work across research, drafting, follow-up, and coordination.
A hypothetical use case would be a product team assigning one agent to gather internal research, another to summarise open tasks, and a third to prepare meeting follow-ups. That does not confirm any specific customer deployment, but it illustrates the kind of workflow Perplexity appears to be targeting.
The competitive angle also matters. The source explicitly compares Teammates with OpenAI’s Dots and Anthropic’s Claude Duet, suggesting that multi-agent workspaces are becoming a key battleground in enterprise AI. For buyers, the real differentiators will likely be reliability, permissions, memory handling, and how well these agents fit into existing business processes.
Źródło: oryginalny artykuł opublikowany przez serwis The Verge. Powyższe stwierdzenia odzwierciedlają informacje zawarte w tym źródle.
Zapoznaj się z powiązanymi artykułami w naszej sekcji AI agents news section.
Partner note — Implementi.ai: As companies test agent-style tools for ongoing work, one immediate challenge is keeping meetings, action items, and follow-ups organised across teams. A focused workflow such as an AI meeting assistant for summaries and next steps can be a more controlled starting point than broad autonomous deployment.
Ta strona używa plików cookie.