AI superworkers for email
Hire a digital employee, give it a mailbox and a mission, and let it read, draft and organise all day. Nothing leaves your outbox until you say so.
No credit card. No autonomous sending in this release.
Counterparty accepted 4 of 6 redlines; the draft answers the last two.
“We’re good on clauses 1–4. Clause 5 (liability cap) and clause 6 (data residency) still need wording we can both live with.”
“Thanks — glad clauses 1–4 are settled. On clause 5 we can accept a 12-month liability cap at 1× fees; on clause 6 we’ll host EU data in-region.”
She asked for 12% off; the draft counters at 8% with a 12-month commit.
A simple request; the draft replies with the invoice attached.
Sender domain does not match the vendor on file. Held, not drafted.
Not because there is too much of it — because the messages that need your judgement arrive in the same pile as the ones that do not.
Once to work out whether it matters, and again to actually deal with it. The sorting is the work, and nothing does it for you.
After a meeting, a flight, or a night's sleep, you re-read an entire inbox to find the three threads that need a decision.
Today's smart replies draft as if they had never met you. You re-explain the same context, to the same tool, every time.
The alternative sends on your behalf with a confidence you never granted, and leaves you no way to check what went out.
Each superworker has an identity, a mission, a memory and a permission profile — and runs one mailbox on your behalf. You manage a workforce, not a feature.
Drafts and recommendations are staged in one review queue. Approve, edit-then-approve, or reject. Nothing reaches a recipient without a human confirming it.
Set a separate authorisation level for replying, organising, writing files and sending attachments — and drop everything back to advice-only with one click.
Every handled message settles into long-term memory and a growing map of people, companies and commitments. You can open that memory, edit it, or delete any entry.
A short daily card from each superworker: yesterday's progress, today's plan, what needs your decision. Status becomes something you read, not something you chase.
Mail, memory and the search index live on your computer. After a one-time sign-up online, the application keeps working offline, and you choose which AI provider ever sees a word of it.
Every action leaves structured residue — people, organisations, facts, commitments, provenance — so the system knows more about your work each month.
Sign in with Google — or look around in Demo mode first, with sample mail and no account. Claim your permanent handle whenever you are ready.
Pick a role template or write your own, give it a mission in one line, set its permissions, and connect a mailbox. Minutes, not an afternoon.
New mail is triaged against its mission and your history. Routine threads get drafted; anything needing judgement is surfaced with a one-line reason.
Approve or edit each draft from one queue. As it matches your style, you widen what it handles — on your schedule, reversible at any point.
Two authorisation tiers ship, and both end with a human. There is no tier that sends on its own.
Analysis and advice. It reads, explains and recommends. It drafts nothing and does nothing.
It drafts and stages. You approve every send. This is the ceiling — autonomous execution is not part of this release.
Join the early-access list and we will write when there is something real to show you.