AI agents can now sign in, pay and work inside your apps for you. So the question moves from "can the AI do it?" to "what did we hand over, and where does a person still say yes?"
Cursor Vault lets agents hold your logins and pay for things inside your apps. What matters now is who holds the keys and where approval sits.
Two launches this week, both from the same company: SpaceX. Grok Bot can now save your logins, cards and contact details in a Vault, then use them to sign in or check out for you. Cursor's agents can now click and type in any app on your computer, not just the browser. In the demo, an agent drives the iOS Simulator to test what it built.
I like how both handle the trust part. The Bot never sees a saved password, card number or security code, and it asks before it uses a card. Cursor's agent looks at the screen to check its own work, instead of waiting for a person to click through.
So this is a policy chat, not a ban. Decide which accounts an agent may use, which steps still need a human yes, and who keeps the record. And if you ship apps, build, run, look and fix just lost its slowest step.
I'm a big Cursor and Grok Bot user, so I'm keen to test this out.
Agents that finish the job are here. Now the real work is deciding what they get to hold.