Key points
- Superhuman, formerly Grammarly, is rolling out a personalized Go assistant that learns context across connected apps and users' writing styles.
- Go adds 20 connected agents through its Go SDK, including DeepL, Legora, Amplitude, Databricks and Blackboard; agents now work natively in Docs.
- Go replaces the AI tools Superhuman Mail used until now, and handles drafts and inbox organization across Google and Outlook accounts.
- CEO Shishir Mehrotra says Go has over 1 million active users.
Why it matters for marketers: Writing, email and doc tools are folding assistants and third-party agents into one layer, so marketers should check how drafts, brand voice and data connections behave there.
Superhuman, the company formerly known as Grammarly, is expanding its Go assistant across its product line. According to TechCrunch, the update lets Go learn context from connected apps, offer more personalized prompt suggestions, and pick up a user's writing style for email and Docs. A revamped onboarding flow asks users to connect their tools and load work context up front.
Go is also becoming proactive. It watches your schedule and email and drafts responses. In Docs, Go agents work natively and can run background tasks, and users can set routines that trigger workflows on events. In Mail, Go takes over from the AI features used until now. Superhuman is adding 20 connected agents through its Go SDK, including DeepL, Legora, Amplitude, Databricks and Blackboard.
For marketers and content teams, the notable shift is that drafting, research and meeting follow-up are being tied together in one assistant. Shared live notetaking, following Superhuman's September acquisition of Fathom, lets Go produce post-meeting outputs such as a first draft of a proposal. Enterprise search over company knowledge bases and connected tools feeds the same context into drafts.
CEO Shishir Mehrotra said Go has more than 1 million active users and, while work-focused, can also help with personal email and calendars. Superhuman wants users to pick Go over assistants such as OpenAI's Dots or Claude Duet. Teams evaluating it should test how well it holds a brand voice, what data each connected agent can reach, and how the proactive drafting is governed before turning it on widely.
Sources
This brief is our own summary. Read the full reporting at the source.