From group thread to todo: the handoff small teams miss
Opitor is an AI-native operating system for small companies, where every person gets their own AI chief of staff. Ask a ten-person company where work gets decided and the answer is usually a group thread. Ask where it gets tracked and the answer is somewhere else, or nowhere.
That gap is the most expensive thing in a small company’s week, and it is not a discipline problem. It is a missing handoff.
Why decisions in a group thread disappear
A group thread is a good place to decide things. It is fast, everybody sees the reasoning, and you can settle something in four messages that would have taken a meeting.
It is a terrible place to keep things. A thread is ordered by time, not by importance, and the message that settled something looks exactly like the fourteen messages around it. An hour later it has scrolled. A day later it is findable only by somebody who already remembers roughly what was said and roughly when.
So the decision does not vanish because people forgot it. It vanishes because nothing was created at the moment it was made. Everybody left the thread with a shared understanding and no artefact, and a shared understanding decays at a different rate in each person’s head.
The failure shows up a week later in one of two shapes: two people each believe the other is doing it, or one person has been doing it and nobody knew.
The question nobody owns
Watch the end of a thread where something got decided. Somebody says “great, let’s do that”. Then the thread ends.
Nobody asked who is writing this down. Not because the team is careless, but because the question has no obvious owner. The person who raised it feels like they have done their part. The person who agreed feels like agreeing was the contribution. Whoever is most senior is usually the busiest, and whoever is most junior does not feel entitled to assign work to anyone.
In a larger company there is a role for this: somebody runs the meeting, takes the notes and sends the actions. At ten people there is no such role, and inventing one is worse than the problem, because appointing a note-taker for a chat thread turns a two-minute exchange into a process.
The answer is not a role. It is making the handoff so small that the person who is already there does it without it feeling like a task.
The smallest handoff that works
Three fields. Who, what, and by when.
Who is one named person, not a team and not a pair. A todo assigned to two people is assigned to nobody, because each of them is reasonably waiting to see whether the other one moves first. If the work genuinely needs two people, it needs two todos with different verbs in them.
What is written so it can be recognised as done. “Look into pricing” cannot be finished. “Write the new pricing page and send it to me” can. The test is whether a stranger could look at it and say yes or no.
By when is a real date, not “this week” and not “soon”. A date that is slightly wrong is still useful, because it is visible and it can be argued with. No date at all is invisible, and invisible work is what gets dropped.
Notice what is not on the list. No priority field, no status workflow, no category, no estimate. Every extra field has to be filled in at the exact moment when somebody’s attention is already leaving the thread, and the handoff only works if it survives that moment.
Where a chief of staff comes in
This is the part that changes with a chief of staff reading along, and it is worth being precise about what it does and does not do.
It reads your messages, so it sees the thread where the decision happened. When something in there looks like a commitment, it proposes a todo. The proposal comes with the three fields already filled in as best it can: the person it thinks agreed, the thing it thinks they agreed to, and the date it thinks was meant.
Then it stops. The proposal is a suggestion, and a person confirms it or does not. That is deliberate: something that silently created todos from your conversations would fill your list with everything you ever discussed, and a list you do not trust is a list you stop reading.
What you get is the hard part done and the judgement left where it belongs. Recognising that a commitment was made and reconstructing who owed what by when is most of the work. Deciding whether it is real takes two seconds.
There is a second effect that is easy to miss. Because the proposal is made in the moment rather than at the end of the week, the details are still recoverable. A week later nobody can reconstruct whether the date was Thursday or the Thursday after.
Reading the proposals you ignored
Once a week, look at the proposals that were made and not confirmed.
This is a small habit with a surprising amount in it. Some of what you ignored was genuinely not a commitment, and that is the system working. But some of it will be things you quietly decided not to do without ever saying so, which is a different situation entirely, because somebody in that thread is probably still waiting.
The pattern across a few weeks is more useful than any single item. If the same kind of proposal keeps getting ignored, either the team keeps making commitments it has no intention of keeping, or there is a category of work that everybody agrees to and nobody owns. Both are worth knowing and neither shows up anywhere else.
Look at it alone, not in a meeting. This is a review of your own week, and turning it into a group exercise turns a useful list into a performance.
Do not turn the thread into a task board
The obvious overcorrection is to make the group thread carry the tracking as well as the deciding: a pinned message with a checklist, or a convention where every commitment gets a formatted reply.
It fails for the same reason the original problem exists. A thread is ordered by time, so the pinned checklist is out of date within a day, and now you have two sources of truth with the stale one in the more prominent position. Meanwhile the formatting convention makes the conversation stiffer, and people start avoiding the thread for exactly the quick exchanges it was good at.
Let the thread be a thread. Deciding happens there, in its own fast and messy register. Tracking happens in a list with due dates, where a deadline can be seen against everything else you owe that week. The handoff between the two is the thing to get right, not the boundary itself.
Where Opitor fits
Opitor keeps both halves in one product and makes the handoff between them the point. Messages, direct and group, sit next to the Todo List and its deadlines, which are laid out on a calendar by due date. Your chief of staff reads your messages, proposes todos from what it sees, and waits: it drafts and you send, everything it writes is signed in purple, and nothing it wants to remember is used until you confirm it. Each person gets their own chief of staff, private messages stay private, and administrators have no override. It is free to start, with no paid plans yet. There is more on what is an AI chief of staff, on Opitor vs Slack, and on the questions and answers page.