If you’ve been keeping an eye on how artificial intelligence is changing the way we work, you’ve probably noticed a common frustration: AI amnesia. Every time you open a new chat window, it’s like meeting a brilliant intern who has completely forgotten everything you worked on together yesterday. You have to re-explain the context, upload the same documents, and remind it of your team’s preferred formatting. It’s exhausting.

Perplexity, a company that has already shaken up the search engine landscape, seems to understand this pain point intimately. In a recent update, they’ve retired their “Spaces” feature and evolved it into something much more ambitious: Projects.

At its core, the transition from Spaces to Projects is about shifting AI from a fleeting conversation partner to a long-term collaborator. Rather than treating each task as a one-off query, Projects act as dedicated hubs for ongoing work. Think of it less like a chat interface and more like a shared virtual war room where you, your team, and Perplexity’s AI can all roll up your sleeves and get things done.

What makes this iteration interesting is the introduction of a persistent, hierarchical file system. You can upload documents, import folders, or connect cloud drives, and the AI actually builds on them over time. Instead of starting from scratch and generating a brand new output every time you prompt it, the AI—which Perplexity refers to as “Computer” in this context—can open a file, read the latest version, make edits, and save those changes right back to the same document. It’s a subtle shift, but one that mirrors how actual human coworkers operate.

The real magic trick here, though, is something Perplexity calls “Brain.” Brain is essentially a self-improving memory system that runs in the background. Between tasks, it studies the shared sessions and files within a Project to maintain a living memory of the work that has been done. When you come back the next day to start a new task, Brain applies what it has learned so you don’t have to spend the first twenty minutes reassembling the context. It can even run automatically on a schedule or on demand to review everything and update the overall Project summary.

Of course, introducing shared AI workspaces immediately raises questions about data privacy and access. Perplexity seems to have threaded the needle here by keeping a strict wall between shared context and personal data. When you invite team members to collaborate on a Project, everyone works from the same shared files and history. However, any local desktop files you connect, or personal API connectors you use (like pulling from your personal Google Drive or Notion), remain entirely scoped to your account. Your teammates can see the output of the work, but they don’t get backdoor access to your private drives or credentials.

They’ve also clearly designed Projects with existing enterprise workflows in mind. You can set Project-level instructions—like a guardrail command to “never publish anything without asking for permission first”—that act as high-level directives for the AI. For more granular, repeatable processes, you can upload “Skills,” which tell the AI exactly how to format or structure specific types of recurring work.

Perhaps the smartest move is how Projects integrate with the tools teams are already using. By typing a simple command in a connected Slack or Microsoft Teams channel, you can bind a conversation thread directly to a Project. The AI uses the channel’s chat history as context, and any tasks kicked off in Slack land right in the Project’s history. It’s a frictionless way to turn casual watercooler brainstorming into actionable, contextualized work.

For existing Perplexity users, the migration is seamless; all old Spaces have automatically been converted to Projects, carrying their files and artifacts with them.

Ultimately, this update signals a maturation in how we are going to interact with AI in the workplace. We are moving past the novelty of the “prompt box” and entering an era of persistent, contextual collaboration. By giving their AI a memory, a shared file system, and a deep integration into existing team communication channels, Perplexity isn’t just giving us a better search tool—they’re trying to build the ultimate digital coworker.