AI has gotten very good at generating answers. The bigger opportunity now is helping people take action.
That shift is already underway, and AI is moving from chat into real workflows: researching, navigating applications, and completing multi-step processes across systems. But the moment AI moves from answering questions to getting things done, one problem becomes impossible to ignore: secure access.
Secure access, in this context, means ensuring the right human or AI agent can reach the right application or credential at the exact moment an action is taken without exposing sensitive data or stopping the workflow to ask someone to log in manually. Every meaningful agentic workflow depends on this, but most existing access protocols weren't designed for it.
That's why we're expanding our partnership with Perplexity, by making 1Password’s secure access capabilities seamlessly integrate with Perplexity Computer.
Perplexity is building an orchestration platform that coordinates models, tools, and connectors to automate complex work. As Perplexity Computer operates across enterprise environments, access stops being a convenience question and becomes a trust question.
Consider what this looks like in practice: It's mid-March at a twelve-person CPA firm, and twenty new clients have dropped off boxes of files. In past years, the junior staff would spend ten days logging into Chase, Fidelity, Coinbase, and a dozen other portals one client at a time. This year, the senior CPA hands the intake list to Perplexity Computer. The agent asks 1Password for each credential, pulls the 1099s, logs into the IRS practitioner portal, files the extensions, and drops everything into UltraTax. The credentials never touch the model; the CPA spends her fourteen hours reviewing returns instead of typing them, and the firm takes on six more clients than last year because the bottleneck moved.
That's what Perplexity Computer does. And 1Password helps execute these workflows quickly and securely.
"We're focused on expanding what AI can do on a user's behalf," said Dmitry Shevelenko, Perplexity's CBO. "To do that effectively in the enterprise, secure and seamless access has to be built into the experience from the start."
AI agents don't behave like humans. They operate probabilistically, persist across workflows, and act at machine speed. A single workflow might touch a browser session, a system login, a token, and a service credential all in sequence, with all requiring different permissions. Without a secure way to provide agents access, you're either slowing the workflow down with manual authentication or accepting credential sprawl as the cost of moving fast. Neither works at scale.
In the next phase of this partnership, we're working toward a model where access is provisioned dynamically as part of the workflow itself. A human stays in control by defining what should happen and what's allowed by the agent executing the work, but every action remains authorized, governed, and auditable. Crucially, credentials are never exposed to models or prompts.
This builds directly on 1Password® Unified Access, our recently released platform for discovering, securing, and auditing access across human, machine, and AI agent identities. Expanding the Perplexity partnership is a direct demonstration of how Unified Access works in practice. "Our partnership with 1Password helps ensure that as we expand what AI can do, we're doing it in a way organizations can trust," Shevelenko added.
The diagram below illustrates how 1Password Unified Access grants secure access to Perplexity's agent, without directly exposing credentials to it.
Security and productivity have long been framed as a tradeoff, but they don't have to be. If AI is going to deliver real value in the enterprise, the secure path has to be the easy path. That is what we're building together with Perplexity.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。