Why a person approves every send

A newsletter goes to people who trusted you with their address. A wrong mail cannot be taken back. So in Geenius Newsletter, nothing goes out until a person says yes.

The yes is tied to what you saw#

An approval is not a switch you flip once. It holds the exact content of the issue and the audience it goes to. If someone changes a sentence or adds a list after you approved, the send needs a new approval. What you approved is what goes out.

Your AI prepares, you decide#

Your own AI can do a lot of the work: start a draft from your notes, fill in sections, check that an issue is ready. What it cannot do is approve, schedule a send or mail a list. Those actions do not exist on the door it uses. When your AI thinks an issue is ready, it asks, and you decide in the app.

Why not let it send?#

Because the cost of a mistake falls on your readers, and on your sending reputation. An AI that misreads a note, or a page that tries to give it instructions, should never be one step from a mail to thousands of people.

What an approval looks like#

The approval card shows the subject, the content as the reader will see it, the lists it goes to and how many active readers that is. Below it are the test you sent to yourself and when. Approve, and the card records who said yes and the exact version of the issue.

If you then fix a typo, the card goes back to waiting. That can feel strict, and it is meant to be: the version that goes out is always a version a person read.

Pausing is part of the plan#

Approval is the first safety net. The second is that sending watches itself. A hard bounce or a complaint suppresses the address, three soft bounces count as a hard one, and when complaints pass 0.1 % of a send the send pauses. You look at what happened, then decide whether to go on.

Where it stands#

Approval is built and in the app. Newsletter is in early access, and sending to real inboxes is not switched on yet: sends run in test mode.