Staying inside platform rules

Every automated message passes through the same gates before it is sent. This page explains them, because "why didn't my automation fire?" is nearly always one of these, and because understanding them is what keeps your account alive.

Gate 1: is this kind of automation allowed here at all?

Each platform has a posture toward automation (allowed, restricted, or prohibited) and a separate answer for whether automated direct messages are permitted at all.

LinkedIn is the clearest example: automated DMs are prohibited there, so a send-DM action on LinkedIn is refused. The builder turns the option off, and the server refuses it independently, because a disabled button is not a control, only a server-side refusal is.

Gate 2: the messaging window

Platforms permit a reply only within a window after the person contacted you. Outside it, the send is blocked.

This is the most common reason a message does not arrive, and it is not something we can configure around. It is also why automations that fire on a fresh interaction (a comment just posted, a story just replied to) work far better than anything that reaches back to an old conversation.

Gate 3: disclosure

Where a platform requires automated messages to be disclosed as such, the disclosure is not optional and cannot be switched off. If an outgoing message does not already contain it, it is added on the send path: not merely requested in a prompt, because an instruction is not an enforcement.

Gate 4: opt-out

Someone who has opted out is not messaged again. This is absolute and it applies across automations and broadcasts alike.

Gate 5: velocity

Sends are paced. A flood of messages from one account in one minute is the signature every platform's abuse system looks for, so the product will not produce it, and a large broadcast drains rather than bursts.

Blocked is not broken

A blocked send appears in automation analytics. A steady number of blocks is normal: it means the gates are catching sends that would otherwise have put your account at risk.

A rising proportion of blocks means the automation is aimed wrongly, usually at people whose window has closed.

What we do not do

• We do not compose outbound text with a model at send time. See automations overview.
• We do not message people who have not interacted with you.
• We do not evade platform limits, and we will not add a setting that does.

Your part

The rules that matter most are the ones no software can enforce for you: deliver what you promised in the post, answer like a person, and stop when someone is not interested.