Perplexity Pushes Multi-Model AI Worker as Competition Heats Up

Perplexity has introduced a new product that moves beyond the typical chatbot format, positioning it as a continuous digital worker designed to operate across tasks with minimal input. The launch reflects a broader shift in how artificial intelligence tools are being framed, from assistants that respond to prompts to systems that can manage ongoing workflows.

The system is described as running on a dedicated machine, such as a Mac mini, and remains active even when the user is offline. Access is designed to be simple, with a keyboard shortcut triggering the interface, but the underlying structure is more complex. Instead of relying on a single model, the platform routes tasks across multiple AI systems at once, assigning different parts of a job to models suited to specific functions.

This multi model approach is becoming more common across the sector, though Perplexity is placing stronger emphasis on parallel execution. Tasks are broken into smaller components and processed simultaneously, which can reduce turnaround time compared with traditional single model responses. The company argues this method allows for faster and more tailored outputs, particularly for research heavy or multi step work.

Accuracy remains a central part of its pitch. Perplexity has long positioned itself as a search driven AI, requiring responses to be grounded in external sources. This differs from systems that generate answers primarily from internal training data. While claims around lower error rates are difficult to verify independently, the focus on citation and retrieval reflects growing demand for more reliable outputs.

The product also leans into automation. It is designed to handle extended workflows over hours or longer periods, stepping in only when user input is required for key decisions. This type of functionality is still evolving across the industry, with many tools facing challenges around consistency and oversight when operating without supervision.

Pricing places the service at a premium tier, with a monthly subscription aimed at professional users. The company suggests the cost can be offset by productivity gains, though such claims tend to vary widely depending on how the tools are used in practice.

Perplexity’s broader strategy includes a publisher programme that compensates media partners when their content is referenced. This approach comes as AI companies face increasing scrutiny over how they use and attribute third party material. By offering payment for citations, the company is attempting to position itself differently within that debate.

Competition in the space remains intense. Established players continue to invest heavily in their own systems, while newer entrants experiment with alternative models and architectures. The idea of a fully autonomous “AI employee” is still at an early stage, and many questions remain around reliability, accountability and long term costs.

For now, Perplexity’s latest offering adds to a growing list of tools that aim to handle more complex tasks with less direct input. Whether it reshapes expectations will depend on how it performs outside controlled demonstrations and how users respond to the balance between capability and control.