The short version
- Every campaign email can carry a short summary an assistant reads on its own: who you are, what you offer, and the record behind each claim.
- When the assistant asks a follow-up, Runner answers only from those records, or says no. Nothing is made up.
- If the assistant says no on its owner's behalf, that no is the same as an unsubscribe, and it sticks.
Your buyer's assistant reads your email before they do
When the person you write to has an AI assistant sorting the inbox, summing up offers and deciding what reaches them, your email is read by that assistant first.
An assistant is good at one thing a busy person is not: checking. Give it something to check, and your email is easier to pass on.
The short version
Write for the person, and hand their assistant the facts behind every claim.
Every email carries a summary the assistant can read on its own
When your campaign's brief cites its sources, each email in it carries a short, machine-readable summary in the open schema.org format: who is writing, what you offer, each claim with the record behind it, and how to say no.
The assistant needs no account and no connection to read it. A person who never looks at it reads exactly the email they would have read anyway.
Follow-up questions get answers from your records, never made up
If the assistant wants to know more, it can ask. Runner answers only with the records your summary cites, with the source beside each answer, or it declines. There is no step where a model writes an answer.
The assistant can ask about that one email and nothing else, so a question can never reach your other customers or campaigns.
Runner speaks A2A, the open standard for agents at different companies, so an agent built on it finds Runner from the email and asks there.
A written rule decides what is shared, and every decision is kept
What Runner may tell an outside assistant is decided by a rule a person can read, by the kind of information. Anything it may not share, it doesn't.
Every decision is recorded, refusals included, so you can always see what was said about your business and what was held back.
AI agents that work togetherWhen the assistant says no, the no sticks
An assistant can decline on its owner's behalf in the way that suits it: a reply, the unsubscribe link, or its own call to Runner. All three do exactly what a person's one-click unsubscribe does.
That person is not emailed again. A no from their assistant counts as theirs.
Questions about selling to AI agents
How do I sell to someone whose AI assistant reads their email?
Give the assistant facts it can check. A Runner campaign email carries a short summary it can read on its own, with the record behind each claim, and answers follow-up questions only from those records.
Does the assistant need to connect to Runner first?
No. The summary rides inside the email in the open schema.org format, so any assistant that reads the email can read it. Asking a follow-up is optional.
Can an assistant get Runner to say something that isn't true?
No. A follow-up is answered only with the records the email's summary cites, or declined. No model writes the answer.
Does Runner speak A2A?
Yes. Runner publishes an A2A Agent Card, so an agent built on the open standard can ask about your email, answered only from its records, or decline it for its owner.
What happens when an assistant says no for its owner?
The same thing as a one-click unsubscribe. That person is not emailed again, whichever way the no arrived.
Give your business a growth team.
Create your account and pick a plan. Your team is built once the plan is paid.
