When the founder is the bottleneck: handing follow-up to each person's chief of staff
Imagine a company of ten people. When someone asks what each person owes the others, the founder is the one most likely to answer without looking anything up. That is a pattern a small company can fall into, not a rule about every founder.
That is not bad management. It is arithmetic. The founder was there for the first version of every conversation the company has ever had, hired each person into it, and for a while genuinely was the fastest place to ask. The habit forms because it works.
Then the company adds a fourth and a fifth and a sixth person, and the habit keeps working right up until the day it does not.
How it shows up in a week
Monday, somebody asks the founder where a thing stands, because asking is faster than finding out. Tuesday, the founder remembers that a supplier was supposed to come back last week and chases it. Wednesday, two people are each waiting on the other and neither of them knows it; the founder spots it at nine in the evening and unblocks it with two messages. Thursday, a date nobody wrote down arrives.
None of those is a crisis on its own. The drag comes from being pulled out of other work repeatedly. Before treating that as a personal discipline problem, list the interruptions and ask which follow-ups have no owner besides the founder.
The symptom to watch for is not that the founder is busy. Everybody is busy. It is that follow-up in this company has exactly one point of failure, and that point is a person who also has other work.
The founder as the company’s short-term memory
The clearest way to describe the situation is that the founder is holding the company’s short-term memory in their head.
The analogy is about dependence, not a specific number of things a person can remember. When a promise lives only in one person’s recall, it can be missed without anyone else knowing it was due. A company running on one person’s recall has that failure shape: everything is fine, everything is fine, and then something that mattered simply never happened and nobody noticed until later.
It also means the company can only move as fast as one person can remember. Every new commitment has to pass through the same small buffer. Hiring somebody to take work off the founder often makes the load worse before it makes it better, because the new person generates commitments the founder now also has to track.
It is fragile in the ordinary way too: when the founder is away, follow-ups with no other owner may wait for their return.
What actually has to move
The instinct is to fix this with a process: a standing meeting, a shared board, a weekly status document. Those help a little and they fail in a predictable way, which is that maintaining them becomes one more thing the founder has to remember to do.
The thing that has to move is smaller and more specific. It is not the work, which is already distributed. It is the three jobs that sit on top of the work:
Knowing what you committed to. Being reminded of it at the point it matters, not after. Writing the message that chases somebody else.
In most small companies those three jobs are centralised by accident, all three on the same person. There is no reason they have to be. Each one is about a single person’s own commitments, and each one can sit with that person.
Each person gets the same equipment
The arrangement that replaces the founder’s memory is not a system somebody administers. It is that every person has the same three things the founder currently provides for them.
Their own todos, with their own dates on them, so that a commitment lives somewhere other than in a conversation that scrolls away.
Their own follow-up, so that a date approaching is their problem before it is anybody else’s, and they hear about it from their own chief of staff rather than from the founder on a Tuesday.
Their own drafting, so that chasing a colleague or a supplier costs thirty seconds rather than the ten minutes of dread that makes people put it off until it is late.
When those three are in place for everybody, the founder’s job changes shape. What is left for them is the quarterly goals, which are genuinely company-level, and the small number of things that are actually stuck. Stuck is a different category from late, and it is the only one that needs the person with the most context.
Two weeks to hand it over
Doing this as an announcement does not work. People have heard announcements about new systems before.
Week one, days one to three: the founder uses it alone. Every commitment you make in a conversation becomes a todo with a date, the same day. You will find out quickly whether you are willing to do this, and if you are not, the rest of the plan is theoretical.
Week one, days four and five: bring in three people. Not the whole company. Pick the three whose work you currently chase most, and ask them for one thing only, which is that anything they promise in a message ends up as a todo with a date on it. Say plainly that the deal is you will stop asking where things stand.
Then keep that promise, which is the hard part. For about a week you will know something is slipping and have to not ask. That week is the whole transition. If you break it, the company learns that the founder is still the follow-up system and goes back to using you as one.
Week two: the rest of the company, with the three early people answering questions instead of you. By the end of the second week the test is simple. Pick something you would previously have chased. Look at whether it moved without you.
Where the line is
Handing follow-up to each person’s chief of staff is not the same as handing over authority, and it is worth being explicit about the limits before somebody assumes them.
AI replies are drafts you choose to send. A reminder you create for yourself can be scheduled to send automatically by Opitor at its due time; it is sent by Opitor, not in your name.
The deployed revision is unknown. Production administrator/non-member requests across all private-message read paths remain to be checked before the founder access boundary can be described as verified.
And it does not decide what matters. Whether a slipping date is a real problem or a date that was always optimistic is a judgement about people and circumstances, and it stays with the person who has to live with the answer.
Where Opitor fits
Opitor is an AI-native operating system for small companies, where every person gets their own AI chief of staff. Each person has their own Todo List with deadlines, laid out on a calendar by due date, and their own chief of staff following those deadlines up and drafting the messages that chase them. The company’s quarterly goals are managed by administrators, which is the part that stays central. Production administrator/non-member checks for private-message access remain pending, and everything a chief of staff writes is signed in purple. The division of labour is set out on what is an AI chief of staff and in the questions and answers.
Updated 2026-09-25