
Perplexity Teammates: AI agents for long tasks, Max only
Perplexity has launched Teammates, AI agents powered by Computer that plan projects, track their progress and report results. Access is currently limited to Perplexity Max subscribers, and no pricing or rollout details were given.
Blog du Modérateur reported on 6 November that Perplexity has introduced Teammates, a set of AI agents built on Computer. According to the report, these agents plan projects, follow their progress and report on their results. The announcement places them as tools for what the source calls long-running tasks, and access is for now restricted to Perplexity Max subscribers.
What the announcement actually commits to
The report describes three capabilities: planning, progress tracking and reporting back. It does not specify which task categories are supported, how a user hands work over to an agent, what level of autonomy is granted, or how results are delivered. No pricing beyond the existing Max subscription is mentioned, and no general-availability date is given. Those gaps matter more than the feature list: the difference between an agent that drafts a plan and one that executes multi-step work without supervision is the difference between a note-taking tool and a workflow replacement.
Context: agentic products across the industry tend to follow a similar path, starting with assisted planning and gradually widening the scope of autonomous action while usage limits and oversight controls evolve. Where a given product sits on that path usually determines whether teams can rely on it for anything with a deadline.
Why this matters for acquisition and monetisation teams
The interesting part for a marketing team is not the agent itself but the reporting loop. Work that spans days, such as competitive research, content briefs, campaign post-mortems or pricing-page reviews, is exactly where human follow-up tends to slip. An agent that keeps a project state and produces a traceable summary could reduce the coordination cost of those tasks.
That said, an agent's output is only as usable as the data it can reach. Most acquisition and monetisation work depends on ad accounts, analytics, CRM records and internal documents. The report does not say which sources Teammates can connect to, so it is impossible to judge today whether it can operate on real campaign data or only on what a user pastes in.
What a team can check this week
- List three recurring multi-step tasks in your acquisition or monetisation workflow and note who currently tracks their status.
- Identify which of those rely on data outside a chat window, since that is where access limits would bite first.
- Decide whether a Max subscription is justified purely by time saved on planning and reporting, independently of the agent's autonomy.
None of these require access to the product. They clarify in advance whether Teammates would change anything in your operations. Until Perplexity publishes details on integrations, execution scope and availability beyond Max, the honest position is to watch the rollout rather than plan around it.
Also worth reading
-
YC startup puts $1M behind five open materials science problems
Discovered Materials, a young YC-backed startup using AI to find chip materials, is offering $200,000 for each of five unsolved materials science problems. The bounty model turns open research questions into a sourcing channel for AI-era discovery work.
-
Axiom Math verifies the Stacks Project with AI agents
Axiom Math's AI agents have formally verified the Stacks Project, a major mathematical library. TechCrunch reports the startup's bet: verification as the answer to unchecked AI-generated proofs.
-
OpenAI IPO talk in 2027: what enterprise buyers should watch
OpenAI's president told TechCrunch the company could be ready to go public in 2027, citing accelerating enterprise revenue. For teams building on AI APIs, the signal matters less as a finance story than as a hint about pricing and priorities.


