Compliance · 6 min read
Life Coaching Inboxes: The Message You Must Not Answer
Life coaching is not therapy. That distinction is real, it matters, and almost nobody arriving in your inbox is holding it.
They found you through a reel about feeling stuck. It was warm and specific and it sounded like it understood them, because that is what good content in this category does. Nothing on the way in told them who you are, what you are trained in, or what happens next. So some of what lands in your DMs will be an enquiry about your programme, and some of it will be a person in real difficulty writing to the first account that seemed to be listening.
You cannot control which arrives. You can control what a machine is allowed to do about it.
The rule that comes before everything else
Never automate a reply to anyone in difficulty. Not a warm one. Not a signposting one. Not one you wrote carefully and read twice.
When a message like that arrives, the only correct behaviour is that nothing automated fires and a person reads it. That is not a trade-off against coverage. It is the requirement the rest of the build is arranged around.
A reply that lands in four seconds does not read as attentive. It reads as: this went somewhere nobody was.
The therapists post works through this posture in full and it is more thorough than this one will be, so read it whether or not you are a clinician. Two things from it carry over exactly, and I will state them briefly rather than restate them.
A negative keyword list is necessary and not sufficient. Distress does not arrive in the vocabulary a filter expects — it arrives as four lowercase words, at 2:40am, in Hinglish, in metaphor, as a reply to a story about something else. And so the posture inverts: build automations that reply to nothing except a short list of situations you decided in advance were safe, rather than to everything except a list of exceptions.
What follows is the part that is specific to coaching rather than to clinical practice.
Your content sets the expectation your inbox inherits
A therapy practice has an easier version of this problem than you do, which is a strange thing to write and is true.
Someone messaging a clinic knows what a clinic is. They have a model of who staffs it and roughly what will happen. Someone messaging a life coach has no such model, and the marketing that brought them in was, by design, emotionally close. The gap between how your content sounds and what your inbox actually is — one person with a phone, running a business — is where the risk lives.
So the boundary has to be stated somewhere people read it before they write, not inside a reply afterwards. Bio, pinned post, the page your link goes to. Plainly: what you do, what you do not do, that Instagram is not monitored continuously, and that it is not a route for urgent help.
Never automate a referral either
There is a tempting middle option and it is worth naming so you do not build it: an automation that recognises a heavy message and replies suggesting the person speak to a professional.
Do not build it. Deciding that somebody needs clinical help is a judgement about a person, and a keyword match is not in a position to make it. It will fire on people who did not need it and stay silent for people who did, and both of those are bad in ways you will only hear about later.
The correct behaviour on an unmatched or uncertain message is silence and a human. That is not a coverage gap. It is the design.
Outcome promises are the second thing a template must never carry
Coaching marketing runs on transformation language, and a template is that language sent in identical words to everyone who typed a keyword, each copy timestamped and unrecallable.
"You will feel clearer in six weeks." "This will change how you show up." "Kitne din mein result?" has no honest template answer, and a generated reply will sound certain because most content in this category sounds certain.
Put the vocabulary on the negative keyword list: guarantee, results, kitne din, how long until, will this fix, will this work for me. Suitability in particular — whether your programme suits this person — is a question about them, the honest answer is sometimes no, and an automated reply says yes every single time.
What claims you may make about outcomes, how you may describe your training, and whether any of it is regulated where you work varies by jurisdiction and by what you call yourself. Take the operational shape from this post. Take the permitted wording from whoever advises you, not from a marketing tool.
The public reply, under a post about being lost
A comment automation replies in public, addressing a handle. Under a post about burnout or feeling stuck, that associates a named person with the subject in front of everyone who follows them, even when the reply itself contains nothing.
Keep the public half completely generic, or leave it out. And do not run comment automations on the posts that invite disclosure — the vulnerable, well-performing ones. Those posts are exactly where an automation is least appropriate and most likely to fire.
What is honestly automatable
A short list, and it is entirely administrative:
- How to enquire. The route, the form, the page. One tap, works on a phone.
- Format and logistics. One-on-one or group, online or in person, how long a session runs, which languages you work in.
- Where to read more. Your site, a recorded talk, a free resource on a specific post.
- Nothing about the sender. If a reply would need to know anything about the person to be correct, it is not a template.
One that catches people: "currently taking new clients" is a fact with an expiry date. It belongs in a temporary automation with a switch-off date written down the day you build it, never in an evergreen template.
The honest ending
Takeover means the moment you reply by hand, the automation stands down in that thread permanently. Use it constantly and early.
No calendar, so nothing books a session. No CRM sync — leads leave as a CSV. No WhatsApp, no sequences, no broadcasts. Nothing can message anyone who has not messaged first. Official Graph API only: seven days on a comment, twenty-four hours in a DM thread, restarted only by their message. That window is a platform constraint and it should never set the pace at which you respond to a person.
Templated replies are free and unlimited; a credit moves only when the AI writes something new. Free tier is 100 credits with no card, packs from ₹499. Cost is not the deciding factor on this kind of account and pretending otherwise would be the wrong note.
Running no automation at all is a legitimate configuration. If the build in the coach build sheet feels too broad for your audience, it probably is, and narrowing it is the right instinct. What the ten checks actually refuse is the other post worth reading before you switch anything on.
