Answer the question you get twenty times a week. Follow up two days later, but only with the people who never replied. No diagram, no flowchart canvas, no code.
Most automation builders hand you an infinite grid and a box of connectors. What a business actually needs is four steps in a row, and the ability to move one up when it is in the wrong place.
Four steps, thirty-seven people partway through.
Thanks for getting in touch! Somebody will be with you shortly.Wait 48 hoursOnly if they have not replied Send — Just checking you got everything you needed.Every step sends through the same path as a message somebody types: the do-not-message list, the block list, the rate limit, the new-contact cap. An automation cannot route around them.
A keyword they text, the first message they ever send you, any new conversation — or somebody being added to a list or given a tag, by hand, by import, or through the API. The last two are how a sequence starts for a customer who has never texted you.
Four kinds of step and no diagram. A message, a wait measured in hours, an "only if" that carries on when it holds, and a stop.
The most useful condition in follow-up: keep going only while the conversation is still one-sided. The moment somebody answers, the sequence gets out of the way.
A run records the step it is on and the time it may next be touched. Restarts lose nobody, and nobody receives a step twice.
A list is a group you send to. A tag is something you know about somebody. A person can be on a list, carry tags, both, or neither — and either one arriving can start a sequence.
The best first automation is the message you have typed out from memory more than ten times.
They text “hours”. You answer with your hours. That is one step and it never needs touching again.
Somebody messages the line for the first time and hears back immediately, even at eleven at night, and a person picks it up in the morning.
Two days later, only if they never replied: one short message. Anybody who answered is already in a conversation and never sees it.
Off stops new people joining and ends everybody already inside it. A flow you switched off does not keep sending for another two days.
Automations cover what happens inside the conversation. When the trigger lives in your own software — a deal stage, a booking, a form — call the API and let Miss Blue handle pacing and delivery.
No. Pick what starts it, add steps, turn it on. The API exists for triggers that live in your own software, but nothing here requires it.
Yes. Put them on a list, or give them a tag — by hand, from an import, or through the API — and a flow attached to that list or tag starts for them. This is the ordinary case for onboarding, a class that starts Monday, or four hundred people from a spreadsheet.
Two things. Everybody on the list who has never messaged that line is a first contact, so they count against the 50-per-line-per-day cap and are spread out — four hundred strangers takes about a week, and the screen says so before you switch it on. And your project's first flow of that kind is read by a person here before it can send, the same as your first announcement.
Nothing, unless you say otherwise. It asks every time whether to start the people already on the list or only the ones who join from now on, and defaults to only new arrivals. The two answers differ by ten thousand people, so it is a decision rather than a setting.
No. Only somebody actually arriving counts as an event — adding a person already on the list does nothing. A second upload of the same four hundred rows starts nobody.
A list is a group you send to. A tag is something true about a person. They are independent: somebody tagged vip may or may not be on the spring promo list, and somebody on that list may carry no tags at all. Either one arriving can start a flow.
No. Every step sends through the same path as a message you type by hand, so the do-not-message list, the block list, the rate limit and the new-contact cap are all evaluated at the moment the step fires. A refusal ends that person's run and records the reason.
The whole message, after normalising case and punctuation. "hours" triggers a rule for hours; "what are your hours?" does not. It is the same rule STOP follows, and it exists so a word in the middle of a sentence never fires something.
That is what "only if they have not replied" is for. The condition is a gate: the sequence carries on while it holds and finishes with a reason when it does not, so a follow-up meant for silent conversations never lands in a live one.
Only if you allow it. A welcome sequence should happen once per person and does by default; a keyword answer should work every time they ask, and you can switch that on per automation.
New people stop joining and everybody already partway through is ended, with a reason on the row. Switching it back on starts fresh for whoever triggers it next rather than resuming a conversation from last week.
Anything up to a month, set in hours. A wait sets a due time and returns; nothing sleeps, so a restart during a two-day wait loses nobody.
Yes, and somebody partway through is not moved by it. If the step they were standing on is removed, their run ends honestly rather than guessing at a replacement.
Build two-way blue-bubble conversations into your product or agent.
Use a complete shared conversation platform without writing code.
Call from your blue line without carrier Spam Likely labels.
Call any dialable number from your Miss Blue business line.