Follow-ups that do not feel like nagging
There are two kinds of small company, and both of them have a follow-up problem.
In the first kind, nobody follows up at all. Things are agreed in a thread on Monday and nobody mentions them again. Some of them happen. The ones that do not are discovered three weeks later, usually by a customer.
In the second kind, one person follows up on everything, all the time, and has become the thing people brace for. Their messages get opened last. The company has solved the dropped-commitment problem and bought a morale problem with the proceeds.
Most founders have been both of these people, often in the same quarter. The swing is not a character flaw. It is what happens when follow-up has no shape, so it alternates between absent and relentless.
What chasing actually is
It helps to be precise about why a follow-up lands badly, because the usual explanations are wrong.
It is not the frequency. A person can be asked about the same thing twice without resenting it. It is not the tone either, though tone gets blamed most often, which is why so many teams pad short messages with softeners that make them longer and more irritating.
What makes a follow-up feel like chasing is vagueness. “Any update on that?” puts the entire burden on the receiver. They have to work out which thing you mean, reconstruct what they said, guess whether you are annoyed, and produce a status report before they can reply at all. That is four jobs handed over in five words.
Compare a follow-up that does the work in advance: the specific commitment, the specific date, and an actual question. The receiver either answers in six words or says what they need. Nothing has been demanded except the thing already agreed.
Precision is the courtesy. Softness without it is not kind, only slower.
The shape of a good one
A follow-up worth sending has four parts and fits in one or two sentences.
It names the specific thing, in the words the person used when they committed to it. Not the category, the thing.
It names the date that was agreed, or says plainly that none was. Many stuck commitments were never given a date, and naming that gap resolves them on the spot.
It offers a way out. “Still Thursday, or has something moved?” is a different message from “Is this done?” The first assumes a world where priorities change; the second assumes a world where people fail. Only one of those is accurate.
And it stops. No paragraph explaining why you are asking, no apology, no restatement of the project’s importance. Length reads as pressure, and explaining why you are entitled to ask is itself a form of pressure.
One sentence, specific, with a door left open. That is the whole craft.
Let it be drafted, and keep the decision
This is where an AI chief of staff earns its place, and it is worth being exact about which part of the job it takes.
Product materials describe a chief of staff that follows up todos and deadlines and can draft a message using the commitment, date and source thread. Its production model context remains unverified. Drafting is the tedious part most often skipped, which is why the message a person sends is often vague.
AI replies stay drafts: you decide whether to send them, and they are signed in purple. A reminder you create for yourself can be scheduled to send automatically by Opitor at its due time; reminders created by somebody else require approval.
The division matters more here than almost anywhere else, because whether to follow up is a judgement with context the system does not have. You know this person shipped something hard last week. You know the deadline is soft because the client moved.
So the useful output is not a sent message. It is a draft on your screen, at a moment when you would otherwise have forgotten, with the decision left to you: send it, change a line, hold it until Thursday, or delete it and go and talk to them instead. None of those is available if nobody remembered in the first place.
Three directions, three registers
The same facts get written differently depending on who is reading, and getting this wrong is what makes follow-up feel graceless.
Upward, to someone senior to you, lead with what you need and why it is blocking. “I need the pricing sign-off to send the proposal; I told them Thursday. Is that still doable, or should I push the date?” You are not chasing a superior; you are reporting a dependency and offering a way to resolve it.
Sideways, to a peer, be short and assume good faith. “Still good for Thursday on the onboarding doc?” They were there; context restated reads as distrust.
Downward, to someone who reports to you, the danger is different. Your follow-up carries weight you may not intend, so name the deadline and hand back control. “Thursday on the supplier list — is that still realistic with everything else this week?” gives them a real option of saying no, which is the only thing that makes yes mean anything.
Same commitment, same date, three messages. The variation is not politeness for its own sake. It is accuracy about what the reader has to do next.
Twice, and then stop
Rhythm is what separates follow-up from pressure, and almost nobody sets one deliberately.
Once before the date, close enough to be useful and far enough out that something can still change. A day or two, usually. This is the one that actually prevents things going wrong, and it is the one teams skip, because before the deadline it does not feel urgent.
Once after, if the date passes without word. Short, not aggrieved.
Then stop. If two specific, dated, easily answerable messages have not produced a reply, the problem is not that the person forgot. A third adds nothing except a record of your irritation. The situation has become a conversation and should be had as one.
Writing the rule down as two is what stops it becoming six. It also removes the guilt that stops it becoming one: you are not hounding anyone by following a stated pattern everybody knows about.
From chasing to being seen
There is one more effect, and it is the reason to bother getting this right.
A team that follows up precisely is a team where commitments are visible. Over a few months that changes what a follow-up means. It stops reading as “I do not trust you” and starts reading as “I know what you took on, and I noticed.”
The people who quietly carry things feel it first. Their work is invisible precisely because it never goes wrong. A follow-up that names the specific thing they committed to is, whatever else it is doing, evidence that somebody was paying attention.
Say the specific thing, name the date, leave the door open, and stop after two. Most of what a small company calls a follow-up problem turns out to have been a precision problem all along.
Where Opitor fits
Opitor’s public materials describe an AI-native operating system for small companies, where every person gets their own AI chief of staff. They describe todos and deadlines laid out on a calendar by due date, with the chief of staff following up and drafting messages; the production model context remains unverified. AI replies are described as drafts you choose to send and sign in purple; a self-created reminder is described as eligible for scheduled sending at its due time. More on the described model is on what is an AI chief of staff, and the questions and answers page states the published sending rule.
Updated 2026-09-25