Three of the six work on a clock and keep going when nobody is watching. One has a single tool. Two work only inside a chat window. Every section below says what runs on its own — and where it stops and asks.
This page is weighted the way the product is. One worker takes most of it, because one worker does most of it.
Scheduled, holding their own state, still working when nobody is signed in.
Answers about a document you hand it, then keeps nothing.
A conversation while the tab is open. Nothing carries on once you close it.
The one job that runs end to end. Give a name and a company, and what follows researches, writes, sends, listens and follows up — pausing at the two points where your judgement is the thing that matters.
At least five web searches before a word is drafted. Then Hunter for a contact address, and Apollo only if Hunter comes back empty, so a lookup that already worked never spends a second credit. If neither resolves one, the prospect stops there and waits rather than going out without a recipient.
The opening email goes out when you release it, and not before. This is the gate, not a setting — there is no path that sends a first email on its own.
Only if nothing has come back. Where a second angle exists, it is picked 50/50 against the first.
Placed only where a phone number is on the record and the sequence is still unanswered. Two minutes at most, recorded, and the transcript is written back against the prospect without anyone typing it up.
Each one checked against the inbox first, so a prospect who has just replied is never sent the next step.
A short note closing the file, and the sequence completes.
Your inbox is checked every five minutes, and again immediately before any follow-up goes out. What happens next depends on what they actually said.
Your booking link goes out straight away.
The return date is read from the reply and the sequence resumes then. Seven days if no date can be found.
The sequence stops.
The sequence stops permanently and the prospect is marked.
A personal reply as well, on top of the booking link that has already gone.
Written against the matching objection script.
A short note asking to be pointed to the right contact.
Interested appears in both columns on purpose. It does two things: the booking link leaves immediately, and the personal reply to that same message waits for you.
These run whether or not anyone is signed in.
Where it keeps going on its own, and where it hands back.
Nothing reaches a new prospect until a human releases it.
With one exception, and it is worth being plain about it: an interested reply sends the booking-link email on its own, before anyone has looked at it. That is the only email sent unprompted, it is a fixed template rather than written prose, and the personal reply to that same message is still drafted for you.
Fifty emails a day by default, counted across every step of every sequence.
Two more run on a clock. Three do not. Each is described by what it actually does rather than by the size of its heading.
A brief every Monday morning, an anomaly check every morning at nine, and questions answered on demand from a rolling thirty-day picture.
The daily check looks for five specific things:
Reply-time figures are withheld until at least five replies exist, so one fast answer cannot present itself as a trend.
Scope. Jordan reads the outbound records Alex produces, and nothing else — not your finance system, not a CRM, no source outside the platform. Nothing here needs approving, because Jordan only reports.
Every Monday at seven, up to three LinkedIn posts are drafted, with topics drawn from the buying signals Alex has recorded during the week.
The detail worth knowing is the name check. Prospects you are actively tracking must not turn up by name in a public post, so generated copy is checked against the watchlist and regenerated up to twice; if a tracked name still appears, the post is refused rather than queued. The same check runs again when you edit a post by hand, because an edit can reintroduce a name the original draft correctly avoided.
The gate. Publishing is a click. The weekly run drafts and queues; it never posts. Scope is LinkedIn posts — ad copy, email campaigns and SEO content are not built.
A conversation with a role-specific specialist that can search the web up to three times per message, see the company context and notes you have written into the workspace, and see facts carried over from your earlier conversations with the team. That is the whole of it.
Neither can:
Sam has one thing Casey does not. Because he writes outreach, his output gets a Review & Send action that opens your connected Gmail with the draft loaded — you type the recipient and you press send. Casey has chat and web search only.
A common misreading. Sam is a customer success specialist you talk to. He is not part of the outbound sequence — the emails that go out on a schedule are researched, written and sent by Alex.
Attach a single PDF, CSV or TXT file up to 4 MB, ask a question, and get an answer from the contents of that document — citing the figures and line items it used, and saying plainly when the document does not contain enough to answer.
Single-shot. Nothing is stored, nothing is scheduled, and each question is independent of the last. Morgan does not remember the previous document or build a picture over time.
The document tool is the extra, not the whole. Like all six workers, Morgan is also a chat specialist — working on figures you paste in, not on a file. Neither surface is connected to an accounting system or a bank feed.
Connecting Gmail takes about a minute. Alex needs nothing else to begin.
Start free trial