Sales automation software that will not send an unchecked claim
Every capability below is listed with the thing it cannot do, on the same row. That is not modesty. A tool that claims everything is exactly the tool a VP Sales refuses to connect to the domain.
Research and writing
The part every tool in the category claims. What differs is what happens to a fact it could not find.
Account research
Reads the sources you connect: CRM records and notes, the account pages you point it at, and the history of previous contact with that person.
The limit. It reads what you connected and nothing else. There is no lead database behind it, no email finder and no enrichment to resell.
First touch drafting
Writes the first email against an angle you choose: new funding, a hiring signal, switching from a competitor, or integration fit.
The limit. An angle is a structure, not a fact. If the funding round is not in a source you connected, the sentence about it goes into the ledger as unverified.
Follow ups that read the thread
Writes each follow up against what actually happened in the thread: what was said, what was ignored, what changed in the record since the last step.
The limit. It does not invent a reason to bump. A step with nothing new to say is flagged as exactly that, and a thread with a reply in it stops.
Your tone, your approved claims
Matches the tone you set and the phrasing it has seen you approve. Approved claims go into a claim library the agent may reuse without re-sourcing them.
The limit. A claim in the library is only as true as the person who approved it. The library records who approved each one and when.
The checks that run before anything sends
Six checks on the draft itself. Each one can hold a send on its own, and each one shows the arithmetic instead of a mood.
Claim ledger
Lists every sentence in the draft that asserts a checkable fact, and marks it cited with the source line it came from, or unverified with a suggested rewrite.
The limit. It classifies assertions, it does not adjudicate truth. A cited claim is traceable to a source you supplied, which is a different and more useful guarantee.
Spam risk scoring
Scores spam risk from 0 to 100 and names every trigger it found: the phrase, the link count, the all caps word, the length, the reading grade.
The limit. It scores the draft. It does not seed test inboxes, check blocklists or tell you which folder a message landed in. Nothing on this site claims otherwise.
Readability and length
Reports the reading grade and the word count next to the range each is judged against, so a rewrite has a target rather than a vibe.
The limit. A readable email is not automatically a good one. This check will pass a well written email that has nothing to say.
Compliance lines
Checks that the unsubscribe line is present, that a postal address block is present, and that the subject line is not deceptive about what is inside.
The limit. These are the mechanical parts of the rules. They are not legal advice and they do not establish that you had a lawful basis to contact the person.
Suppression check
Checks the address against your suppression list at draft time, so a suppressed contact never reaches a queue in the first place.
The limit. It can only suppress what it knows about. Opt outs that reach you elsewhere have to get into the list, by import or by API.
The verdict
Gives the whole draft one verdict: pass, review or block, with the reason line that produced it.
The limit. A pass means it cleared the checks, not that it will get a reply. No tool can score that and this one does not pretend to.
Controls for the person who signs
The section a VP Sales or a RevOps lead reads first, because these are the answers that decide whether the thing gets connected to the domain at all.
Hard send caps
A hard daily ceiling per mailbox, typed by you and enforced by the queue. The agent has no override, no burst mode and no catch-up for a quiet day.
The limit. A cap protects your sending reputation. It does not make an unwanted email wanted, and it is not a substitute for a list you are entitled to contact.
Review queue and approvals
A review queue with three settings: approve everything, approve only what the ledger flagged, or spot check. On Enterprise the setting can be enforced by role.
The limit. Approve-everything is the honest default for the first weeks. Anyone who tells you an AI SDR is unattended on day one is describing a demo.
Audit trail and export
Stores the draft, the ledger, the verdict and the source line behind each cited claim, as they stood at send time, and exports the lot.
The limit. Export windows differ by plan: 30 days on Solo, 12 months on Team, scheduled and unlimited on Scale.
CRM sync and deal notes
Writes the activity back to your CRM: the send, the reply, the deal note drafted from the thread and the next step it suggests.
The limit. HubSpot and Pipedrive on every paid plan, Salesforce from Team. Custom objects and sandboxes are an Enterprise conversation.
SSO, roles and a DPA
Roles and approval policies, SSO and SAML, SCIM provisioning, a signed DPA and a security review, on the Enterprise plan.
The limit. Enterprise has no published price, on purpose. It is scoped per domain and per policy, and it is invoiced.
What all of that produces
One surface, on every draft: the verdict, the four numbers behind it, and the ledger of what the email asserts. This one is a stored sample. The live one is on the homepage.
Spam risk 20/100 Grade 7.3 87 words 1 of 40 today
Cleared every check. Queued against your cap rather than sent immediately.
Claim ledger
2 of 2 cited
-
You opened a second warehouse in Reno in March.
CITED
Source: the press page you connected.
-
Your team still keys returns into two systems.
CITED
Source: CRM note from the discovery call on 2 March.
Why the verdict is first
Because it is the thing being sold. Most tools in this category return an email and nothing else, so the buyer has no way to tell a good draft from a confident one.
The numbers are computed from the text that came back, not from a stored animation. Run the demo twice with different input and every one of them changes.
spam score checker, explained line by line
ai email writer, the drafting side on its own
AI SDR software pricing, flat rather than per seat
Built for the job you came here to do
The same product, entered from the job you came here to do. Each page states its own limits rather than repeating the pitch.
-
AI email writer
The drafting side on its own page: what it writes, and what it refuses to assert without a source.
-
Spam score checker
What the score is made of, line by line, and the two things it deliberately does not test.
-
Email outreach platform
The cap, the queue and the suppression list, which is the part of a platform nobody demos.
-
CRM automation
Activity logging, deal notes and next steps, written from the thread instead of from memory.
-
Prospecting tools
Where a research tool ends and a list vendor begins, and which one this is.
-
Cold email software
The four things a cold email has to get right before volume is even worth discussing.
-
Sales automation tools
What to automate, what to review, and what a person still has to decide.
-
Email sequence software
Follow ups written against the thread, not against a template with a merge field in it.
Every feature here runs in the demo.
Paste two lines about an account and read what comes back: the draft, the ledger, the spam score and the verdict. Nothing is connected to your domain to try it.
Your data stays yours. Prospect data is never used to train a model. Nothing is charged today.