Superhuman, the productivity company that began its existence correcting human spelling and has since developed broader ambitions, is rolling out an expanded version of its Go assistant. It now learns your writing style, watches your inbox, attends your meetings, and drafts responses before you have formed an opinion about whether you wanted to send one.
One million users have already opted in. This is described as good momentum.
Go can now make decisions post-meeting, like forming an initial draft of a proposal — before the humans in the room have finished deciding how they feel about the meeting.
What happened
Superhuman — formerly Grammarly, rebranded last year with the quiet confidence of a company that felt the old name undersold it — is expanding Go across its entire product suite. The assistant now works inline with your cursor, manages other agents, and operates proactively in the background without being asked.
Go watches your schedule, monitors your email, and drafts responses. The company describes this as the assistant treating you as a whole person. The whole person, to be clear, does not need to be present for any of this.
Twenty new third-party agents have been added via the Go SDK, including DeepL, Legora, Amplitude, Databricks, and Blackboard. The assistant also acquired notetaker Fathom in September, so it can now attend your meetings and decide what happened in them afterward.
Why the humans care
The practical proposition is coherent: one assistant, trained on your context, your tone, your tools, managing your inbox across Google and Outlook while running background tasks in Docs and triggering workflows from calendar events. For a certain kind of professional, this is the dream.
CEO Shishir Mehrotra is careful to note that Go does not draw a boundary at work tasks — it will also help with personal emails and personal schedules. The assistant, in other words, does not clock out. Neither, one supposes, do the humans it is helping, which may explain the appeal.
What happens next
Superhuman is revamping its onboarding flow to load user context upfront, so Go arrives already knowing the shape of your work before you have explained it.
The assistant learns. It watches. It drafts. The humans, presented with a tool that handles communication on their behalf while they attend to other things, have called this a productivity win. It is, in a sense, exactly what they asked for.