Chapter 13. Work that finds its way

A working draft, November 2026. This book is being written while The Grove is being built; expect it to change. Checked against the build as it stood on 1 November 2026.


In this chapter

  • What handing out work costs the manager, the business and the people given it
  • Why work belongs to a branch, and nobody can give work to anyone else
  • How everything comes in through one door and is routed in a fixed order, with guesses marked
  • How connectors and the organisation's own website, the Canopy, carry work in and replies out
  • Why work weighs more or less depending on where it came from
  • How playbooks and new structure grow from the work, and what's left out on purpose

Here's a doubt that sounds like plain common sense.

Someone has to hand out the work. An email arrives from a client. A supplier sends an invoice. A customer complains on the website. Somebody has to read it, decide whose it is, and pass it on. Take away the person who does that and the work piles up in a shared inbox until someone gives in and deals with it, usually the most conscientious person in the room, who then quietly becomes the boss anyway.

It's a fair point. Work doesn't sort itself. If nobody is giving it out, something else has to be doing that job, and it had better be good at it.

This chapter is about what that something is.

What handing out costs

In most organisations, handing out work is the manager's job, and it's a bigger job than it looks. It means reading everything that comes in. It means knowing who's busy and who isn't, who's good at what, who's on holiday next week. It means chasing what was handed out to see whether it got done.

A whole industry of tools has grown up to support this. Task trackers, priority fields, sprints and backlogs, capacity plans, timesheets, status reports that travel upward each Friday. They're often well made. But look at what most of them are for. They exist so that one person can hand out other people's time and then check on it.

That has a cost for the business. The manager becomes the place where work queues. When they're away, things stall. When they misjudge who should do something, it gets done badly or twice. And because work belongs to whoever it was given to, it leaves with them when they leave.

It has a cost for people too. Work that's handed to you is work you didn't choose. You can do it well, but it's hard to care about it the way you care about something you stepped towards. And the person handing it out carries a load nobody can carry for long: holding everyone's week in their head.

The work already knows where it goes

As Chapter 4 described, in 1959 the French biologist Pierre-Paul Grassé noticed that no termite directs the building of a mound.1 Each one responds to what the last one left behind. He called it stigmergy: coordination through traces in the work itself, rather than through instructions from someone.

We're not building mounds. But the idea points at something useful. Most work, when it arrives, already carries clues about where it belongs. An invoice says it's about money. A booking says it's about the calendar. A complaint names a product. If an organisation has written down, clearly, what each of its parts answers for, then much of the sorting a manager does in their head can be done by matching the work against those answers.

In 1999 Stephan Haeckel argued that organisations in an unpredictable world need to stop planning and start sensing and responding, with each part clear about what it's accountable for and free to answer in its own way.2 That's close to the shift here. The question stops being who should I give this to? and becomes which part of us answers for this?

Work belongs to a branch

In an organisation on The Grove, work belongs to a branch, never to a person. Nobody can give work to anyone else. People take it. The only exceptions are the system's own: work arriving on a branch with one holder lands with them, and a promise made in a reply lands with whoever sent it, if they hold that branch.

That's the whole of it, and it changes a great deal. As Chapter 7 described, people hold seats in branches because they chose to. When work arrives at a branch, whoever holds it can pick it up. Taking a step on a piece of work is a way of saying I can.

One door

Work reaches an organisation in many ways: email and other connected services, the website, chat between branches, the Field, decisions, things set to happen on a schedule. All of them come in through the same door, and all are routed the same way. We test for this. If anyone adds a new way in that slips past the door, a test fails.

Behind the door, the routing follows a fixed order. First, it looks for an accountability that names the work, on any branch, anywhere in the organisation. An accountability is one of the things a branch has said it answers for, like bookings or supplier invoices. If one fits, the work goes there.

If nothing names it, the router looks for a branch the work plainly belongs to anyway. It waits there as work with no job yet, and its presence tells the branch it may need a new accountability.

If no branch is a plain fit, Wise-Eye makes its best guess, and marks it as a guess. The branch's people can confirm it or move it on, and each move teaches Wise-Eye where that kind of work belongs. Nothing goes out to anyone while the work is still a guess, and Wise-Eye never guesses at high stakes.

And if nothing anywhere comes close, the work goes straight to the root as work no job covers. It becomes a question for the forum: who are we, if this keeps arriving?

Ways in and out

The door has a front side. Before work can be routed it has to reach the organisation, and a reply has to find its way back out. That happens in two ways.

The first is the connector: a link to one outside service, saying exactly what the organisation may do there and nothing more. Four are built. A Gmail mailbox, which can read, label and send, and can never delete anything for good. A Google Drive, which reaches only the files and folders someone picked in Google's own chooser, and only to read them. The organisation's social accounts, through one service that gathers many platforms, so comments and messages come in and answers and posts go out. And its books in Xero, read only, so the money can be seen without anyone typing it in. There's no calendar among them, because organisations on The Grove keep no calendar of people.

A connector belongs to a branch. Whoever holds that branch's seat connects it, and says whose account it is: the branch's own, or their personal one carrying the branch's work. A personal one is flagged. It pauses when they leave the seat, and only they can hand it to the branch. As Chapter 9 described, a mailbox is one of a branch's domains. Any branch whose work it is can answer and write through it, under its own name, unless a proposal has locked that domain to its holders. Reading all of it, setting its rules and removing it stay with whoever holds the domain. Connecting, pausing and removing each go on the trail.

The keys never come inside. Signing in happens on a thin relay outside, which seals the key so that only that one connection can open it. The system stores the sealed form and never the key. The relay can open it only while a job for that connection is in hand, and forgets it within minutes. Remove a connector and its key is gone for good.

A mailbox is looked at every minute or so, and its first sort happens where the mail is: the branch's own rules first, then rules it accepted from Wise-Eye, then Wise-Eye's reading of the sender, the subject and the opening lines. Only work is brought in whole. What's unclear comes in as headers only, with a question, is this ours?, which is let go after a fortnight unanswered. Everything else stays in the mailbox, and all that's kept of it is a bare reference, for a week. A new rule that brings mail in looks back thirty days for what it would have caught. Comments and messages from the social accounts have no first sort: they all come in, and triage files what's noise.

The second way is the Canopy, the organisation's website. It isn't a connector. The system serves it itself, at the organisation's own address and at up to five domains of its own. Its structure, each section's words and its look are held as separate domains, and nothing reaches the open web until the people holding a part publish it. Wise-Eye can draft every word and publish none. The page runs no outside script and sets no cookie, and visits are kept as daily totals: views, never people. A message through its form, or anything another website posts to an address the organisation has opened for it, becomes work. And what it chooses to publish as a feed, any website can read.

Whichever way it arrives, work then comes through the one door like everything else.

When the stakes are high

Some work can't wait and can't be got wrong. An angry client. A letter from a tax authority. That kind of work is still routed, but it's held for whoever holds the branch's seat, with a careful reply drafted beside it. Only a person sends it.

What people can do with work

Once work has landed, the people holding a branch can route it within the branch, refuse it, in which case it rises to the branch above, or say it isn't work for the organisation at all. Any member can usher it to another branch.

Work doesn't sit forever. Work from outside that nobody has taken after two weeks can be taken by the people on the branch above. Work that still has no job after two weeks moves up a branch. If it reaches the root and still no job covers it, it rests there as homeless. That isn't a failure of anyone. It's a sign that an accountability may be missing.

Where it came from matters

Not all work weighs the same, and how much it weighs depends on where it came from, not on anyone's opinion of it.

Some work comes from outside: somewhere along its chain, it reaches something that arrived from a client, a supplier, the world. Some is held: the organisation's own committed work, like a regular cadence, a measure it watches, or a decision it made. Some is going out: a post or a probe with nothing inbound behind it. And some is made here, with nothing felt behind it at all.

Only work from outside and held work make a branch want to grow. Work from outside counts only if it cost the sender at least a real ask: a like or a passing comment shows attention, not need. And demand for something beside the organisation's purpose never grows anything on its own. Made-here work with no job waits folded until someone gives it a job or lets it go.

We chose this after seeing something uncomfortable. In a brand new organisation, almost all the work is imagined. Nobody outside has asked for anything yet. Without weights, the person with the most ideas would generate the most work, and the structure would grow around them. The loudest visionary would build the organisation. Weighting by where work came from lets the world, and the organisation's own promises, shape it instead.

Work is shown in order of how much it matters that someone picks it up. It's never given a score. Why it sits where it does is one tap away.

Four springs, and playbooks

Work rises from four springs. There's committed work: what a branch has said it answers for, some of it set to happen every so often, at set times, on a date, when a number crosses a line, or when a record changes. There's floating work, named in the moment. There's homeless work. And there's work born from decisions, including a proposal's own date for review.

How an accountability's work gets done can be written down as a playbook. When work is routed to that accountability, the playbook starts. Anything it would send waits for a person's approval. After five low-stakes drafts in a row go out exactly as written, Wise-Eye offers to send that kind itself, and a person says yes or no. Anything with more at stake always waits for a person. When people correct what it did, those corrections become suggestions, and the steward accepts or dismisses each one.

Structure grows from the work

When the same kind of work keeps being carried by hand, the branch is offered help: records to keep it in, a board, a practice, an invitation for someone to join the crew. When weighted work keeps arriving with no accountability to name it, one is offered. The organisation's shape follows what it actually does.

Each branch has a board of next moves, seen only by whoever holds it. The cards point to the work; they never copy it.

And where branches overlap, for instance when the same outside person turns up in two branches' work, or promises to one person are split across branches, that shows as a structural tension for people to look at. We decided this on 1 November, and it hasn't been tried with real organisations yet.

What isn't here, on purpose

There's no assigning people. No capacity planning, no smoothing out anyone's load, no timesheets, no priority fields, no sprints or backlogs, no status reports travelling upward, no alerts. When nobody hands out other people's time, that whole layer has no job to do.

What remains is small. The promise, which Chapter 15 explores. The route. The automation. And three things left for a person: to decide, to promise, and to do something on a particular day.

How we decided

The routing order came first: an accountability that names the work, then a branch it plainly belongs to, then the best fit, then learning from where people usher it. That order was settled in October 2026.

The harder question was what to do with work that has no exact home. One option was to refuse it, and send it back to the root every time until someone wrote an accountability for it. That would be clean and honest, and nearly unbearable for a small organisation whose accountabilities are still thin. We chose best guess plus learning instead, with the guess clearly marked and nothing sent while it's a guess. We're not certain the guessing will hold up across many organisations. It's been tested in simulation, not yet with real people.

Simulation has already taught us something. In one simulated world, a small practice of facilitators, every booking landed in the root's inbox first and was then passed on to the new bookings branch, thirty-one times over. The reading took all that sorting as work bouncing between two branches, and six days after the bookings branch was made, suggested merging them back together. The fix was simple once we saw it: the door's first sorting never counts as a pass. But it was a good reminder that the reading only sees what it's told to see.

Choosing to weight work by where it came from followed from watching new organisations fill up with imagined work.

And on 29 October we decided, firmly, not to become a work management tool. There are no reminders sent out into people's other apps and no calendar of people. David Allen's Getting Things Done helps a person keep their own list.3 Kanban, as David Anderson set it out, helps a team see and limit the work in flight.4 Both are good at what they do, and people may well keep using them. Organisations on The Grove aim at something different: the route by which work reaches the part of the organisation that answers for it. Whether people will miss the familiar layer of tasks and priorities, or feel lighter without it, is something we'll only learn when real organisations use it.

What you might take from this

Handing out work feels necessary because, in most organisations, nobody has written down clearly enough which part answers for what. Once that's written, most of the sorting can follow from it, with people confirming and correcting rather than allocating.

Nobody needs to hand out the work: it can find the part of the organisation that answers for it.


Notes and further reading

1

Grassé, P.-P. (1959), "La reconstruction du nid et les coordinations interindividuelles chez Bellicositermes natalensis et Cubitermes sp. La théorie de la stigmergie", Insectes Sociaux. The paper in which the term stigmergy was introduced.

2

Haeckel, S. H. (1999), Adaptive Enterprise: Creating and Leading Sense-and-Respond Organizations, Harvard Business School Press.

3

Allen, D. (2001), Getting Things Done: The Art of Stress-Free Productivity, Viking.

4

Anderson, D. J. (2010), Kanban: Successful Evolutionary Change for Your Technology Business, Blue Hole Press.