Agreement$48,000 at stake
Acme's renewal terms don't match what you quotedTheir email says Net-60. Your signed agreement says Net-30. A corrected reply is drafted.Acme is your largest active accountReview TaskFieldAI for sales leaders
The deal was fine in August.
Nothing in your pipeline review said otherwise. The champion was replying, the notes looked current, the stage had not moved. What had changed was in the emails, and none of it was in the CRM.
One-click Google sign-in · Personal email excluded in code · Export or delete everything anytime
A forecast is a set of claims about relationships nobody has measured. The stage field says what a rep believed on the day they last touched it, and the deal that slips in Q4 stopped being real in Q2.
The problem isn't rep discipline Reps update the CRM before pipeline review, from memory, under time pressure — which is exactly the condition under which people report what they hope is true. More fields will not fix a system whose inputs are recollections.
The seven questions you already ask.
You ask most of these every week, usually without getting an answer to any of them. Here is what comes back instead.
- Which deals are actually moving?Movement measured from the correspondence rather than the stage field: who is replying, how fast, at what length, and whether they are still the one starting threads.
- Did the champion go quiet?Each contact is measured against their own baseline, so a three-day reply is a signal for someone whose median is four hours and nothing at all for someone whose median is three days.
- What did we actually promise to close this?Concessions made in email — a discount, a date, an extra seat, a bespoke SLA — are captured as commitments with the sentence that created them, rather than living in the rep's head until renewal.
- Who hasn't been followed up?Every request made of your team is tracked with who made it and how long it has sat, including the ones buried three replies into a thread about scheduling.
- What changed on this account since last week?Not what arrived — what became different. A date that moved, a term restated, a new stakeholder added to the thread, a champion who stopped asking questions.
- What do I need before this call?An hour beforehand: the history, open commitments on both sides, what changed since you last spoke, and anything unresolved — assembled from your own record of the people in the room.
- Which deal is about to slip, and why?When several signals decline together across multiple exchanges, that is surfaced with the evidence behind each one — so the answer to “why” is a list of emails rather than a hunch.
TaskFieldPipeline · 3 need youExample workspace · illustrative data
Three things changed on live deals since Friday.
Relationship
Marcus has stopped driving the Northwind threadReply time 4 hours → 3 days. Message length 180 words → 24. No questions asked in three exchanges.12 days since a meaningful replyOpen Opportunity
James Wilson is ready to moveHe replied in 9 minutes and asked for the contract. Relationship is 91 and rising.Fastest reply he has ever sent youOpen Ranked by consequence rather than by stage or by arrival. Each card carries why it is here and what it holds up — a list without those is just a differently-sorted inbox.
It will not speak for you uninvited.
It starts by drafting only. Autonomy is something you grant later, in steps, and can take back.
A prospect emailDrafted in the rep's own voice, built from the last twenty emails they actually sent that person — not from a tone setting. Nothing goes out until you allow it to.
A thread with a risk on itFrozen. An active agreement flag stops autonomous sending on that thread entirely, the check runs before any AI call, and there is no setting that turns it off.
Personal emailExcluded in code rather than by a preference. Structurally out of reach, including from us.
When it isn't sureIt says so. A claim it cannot verify comes back unverifiable rather than fine, and a contact without enough history reads “not enough history”, never “healthy”. The full brief →
Straight answers.
Does it sync with Salesforce or HubSpot?
Not yet. Gmail and Google Calendar are the live connectors today; HubSpot and Salesforce are on the roadmap and are not something you would be paying for now. What works today reads your email and calendar — which is where the signals actually are, but it is not CRM sync and we are not going to call it that.
Is this a team product?
Single-user is what ships today. Each person connects their own mailbox and gets their own workspace. Shared team memory and manager-level rollups are roadmap items, so a pipeline-wide dashboard across your reps is not something this does yet.
Will it email prospects on its own?
Only if you grant that, per category, after watching it draft. Even then every send holds for five minutes, respects working hours, and is blocked outright on any thread carrying an active agreement flag.
How long before the signals mean anything?
It reads the last 90 days on connection, and relationship baselines need roughly fifteen interactions with a contact before they are worth acting on. Until then a contact reads “not enough history” rather than being given a score it has not earned.
The signals were in the emails the whole time.
Connect Gmail and it reads your last 90 days of them. The first thing it shows you is what it found — with the sources.
See what it findsNo credit card · Your data exports or deletes in one click