Skip to main content
Most of what Superhawk produces, it produces without being asked. Understanding what runs on its own, and where you sit in the loop, is worth five minutes.

​
The shape of it

Superhawk continuously reads your connected sources, extracts what happened, judges what it means, and drafts the work that follows. You decide what actually happens.

​
What runs on its own

You do not trigger these. There is no “run analysis” button to remember, and no account that goes stale because nobody pressed it.

​
What you approve

Anything that leaves Superhawk:
  • Emails to customers — drafted, never sent
  • CRM writes — drafted, never written
  • Documents shared externally — generated, reviewed by you
Superhawk drafts and you decide. Nothing reaches a customer or your CRM without a person approving it.

​
What it never does

  • Move data between workspaces. Every judgement, memory, and answer is scoped to one workspace, and within it to one account.
  • Send anything to a customer on its own.
  • Change your CRM on its own.
  • Delete your data. Disconnecting an integration stops syncing; it does not remove what was already synced.

​
Everything is evidence-backed

Every health band, signal, brief claim, and chat answer carries what it was built from. This is deliberate and it is the most important thing to know about using Superhawk well: you are never asked to trust an assertion you cannot check. When something looks wrong, follow the evidence. It will usually tell you whether the reading is wrong or your expectation was.

What Runs Automatically

The schedules, and what to expect when.

Accuracy & Limits

Where Superhawk is strong, where it is weak, and how to tell.