Answers you can check
Updated . The permission explanation has been updated to include deliberate sharing into Spaces and access checks when citations are opened.
Most AI answers end with a shrug: trust me. For a draft poem that’s fine. For a contract renewal date, a policy threshold, or a client commitment, it’s unusable. Someone on your team has to verify before they act, and if verification means re-searching five tools by hand, the agent saved no one any time.
In Pearl, cited claims link to the source they came from (the document, the email, the page) and open it for inspection. The answer travels with its evidence, so checking a fact is a click, not a project.
Proof has to survive change. Sources move, permissions shift, documents get replaced. When someone opens a citation, access is checked again. Removed or inaccessible sources are unavailable rather than silently exposed.
Permissions are part of proof, too. Source permissions and deliberate sharing determine what a person and the agent can retrieve through Pearl. An authorized person can share source content into a Space for its members to read. Sharing does not override an explicit denial of access at the source, and retrieval checks the caller’s access.
This is what we mean by answers with proof: not a smarter-sounding agent, but one whose work holds up when someone careful checks it. That bar, whether a careful person would accept it, is the one we design to.