Compound.

Compound is the sixth stage of the Sequence and the one that closes the loop. Signal found the constraint. Source mapped what the work required. Design wrote down who does what. Build produced the spec and the workflow. Deliver shipped it inside the company and measured it against the cost Signal named. Compound is what happens after the win is documented — the stage that turns a delivered sprint into the foundation of the next one.

This is the stage the word in our company name points at. The brand is Compound because what the work produces is compounding — not as metaphor, but as the structural property of running the Sequence on a cadence. Without this stage, you have shipped a project. With it, you have built infrastructure.

What Compound produces.

Compound has three deliverables, and they ship together.

The first is an updated Signal Backlog — the running, prioritized list of validated constraints the organization is working through. Signal generated the constraint that drove this sprint; the sprint generated new information about the rest of the backlog. Compound is where the backlog gets re-ranked against what you now know.

The second is an updated Hybrid Accountability Chart. The chart had one new row when Design produced it for this sprint’s accountability. By the end of Deliver that row was real — a named agent team, a named human supervisor, a position on the AI-assisted-to-automated spectrum, and a documented outcome against the constraint. Compound is where that row lands permanently in the chart and where any neighboring rows get reviewed for the second-order effects of the sprint that just shipped.

The third is a commitment to the next sprint’s focus — the constraint, the accountability, the human owner. Not “we will start the next sprint sometime.” A named next constraint, taken from the re-ranked backlog, with the Human Orchestrator already assigned.

If a sprint ends and the backlog is not updated, the chart has not gained a permanent row, and no next constraint is named, the sprint did not finish. It stopped. Those are different states. Compound is the discipline of finishing.

The two instruments.

Compound has two Skills Library instruments: Sprint Retrospective and Constraint Re-rank.

Sprint Retrospective is the structured review of the sprint that just shipped. It asks the same questions every time. What worked — the specific design decisions that paid off, not “the team did great.” What did not work — the handoffs that broke, the inputs that turned out dirtier than Source claimed, the supervisor reviews that got skipped because the cadence was wrong. What one design change would have made this sprint better — the disciplined version of the lesson, written as a future-tense instruction the next sprint can use. The output of Sprint Retrospective is not a document filed in a drawer. It is a small number of design changes that get installed into the operating system before the next sprint begins. Lessons that do not change the design are not lessons. They are notes.

Constraint Re-rank is the corresponding instrument on the strategic side. The Signal Backlog that existed at the start of the sprint was built on the information the company had at that moment. The sprint produced new information — about what the data really looks like, which accountabilities have unexpected friction, which constraints turned out to be symptoms of a deeper constraint. Constraint Re-rank takes the backlog into that new information and reorders it. Some constraints move up. Some move down. Some get resolved entirely, because the sprint that just shipped quietly absorbed them. Occasionally, a constraint that was not on the backlog at all emerges as the next most important — the kind of insight you only get from operators who have just finished a sprint.

The two instruments run together. Sprint Retrospective looks backward at the work that shipped. Constraint Re-rank looks forward at the work that comes next. Both produce changes to artifacts the company already maintains — the chart, the backlog, the design rules — rather than artifacts the company has to remember to consult.

The Compound Agent.

Compound provides a Compound Agent — the sixth and final coaching agent on the Compound Bench. You run it at the end of every sprint, and its job is to make Sprint Retrospective and Constraint Re-rank impossible to skip.

The Compound Agent takes the documented sprint outcome from Deliver and measures it against the original constraint cost from Signal — the dollar figure or hour figure that opened the sprint. It does not let the team round the number, narrate around it, or convert it into a softer metric. The quoting cycle that took three days was supposed to take four hours; either it does or it does not, and the gap is the data. The agent then runs the Sprint Retrospective conversation with the leadership team — structured, fast, the same questions every time — and produces the disciplined list of design changes that come out of it. It runs Constraint Re-rank against the existing Signal Backlog and surfaces the candidates for the next sprint.

What the Compound Agent does not do is decide what to work on next. The next constraint is the Human Orchestrator’s call — the same call Signal demanded at the beginning of this sprint, made now with one more sprint’s worth of evidence. The agent’s job is to make sure that call gets made with the new information, not with the same map the company had eight weeks ago.

The Compound Agent is Compound’s accumulated reps on this specific conversation, run with the speed and consistency of a coach who has finished a hundred sprints. The team owns the decisions. The agent makes sure none of the questions get skipped.

The one design change rule.

After every sprint, the Human Orchestrator answers one question: what one design change would make the next sprint better?

One. Not five. Not a list of improvements. One.

The temptation at the end of a sprint that produced a real outcome is to list every observation the team made and call the list “lessons learned.” That list never lands. It is too long to install, too unprioritized to argue about, and too vague to enforce. By the time the next sprint begins, the lessons file is closed and the same mistakes are being made in slightly different colors.

The discipline of one design change defeats that failure mode. The Human Orchestrator looks across everything the sprint surfaced — the Source phase that ran too thin, the supervisor cadence that was wrong, the handoff between the agent team and the human reviewer that broke once, the input format that turned out to need cleaning nobody scoped — and names the single change that will most improve the next sprint. That change gets installed before the next sprint begins. It becomes a piece of how the company operates, not a note in a doc.

One design change per sprint is small. Eight sprints of one good design change each is substantial — that is the operating system getting deliberately better at running the Sequence, with the improvements compounding the same way a well-run quarterly Rock produces a company that executes a year later. Disciplined improvement over many sprints produces an agent team dramatically more capable after eight sprints than after one. Undisciplined improvement produces the same agent team it always was, with longer meeting notes.

Infrastructure compounds.

This is the structural payoff of the whole Sequence and the answer to the most important question this book has to answer: why does this work get cheaper and better with each sprint, when most organizational initiatives get more expensive and worse?

Because the artifacts the Sequence produces are permanent, and they accumulate.

Sprint one’s Source phase produced a Knowledge Map for the client delivery function — three weeks of mapping what the function actually knows, where the data lives, what is undocumented, what depends on which person’s head. That Knowledge Map does not disappear when the sprint ends. It lives inside the company. Sprint six in the same function inherits it. The Source phase that took three weeks the first time takes three days the sixth time — not because Source got faster as a stage, but because the map already exists and the work is mostly updating the edges.

The same is true of every artifact in the Sequence. The Hybrid Accountability Chart has one row after sprint one and a dozen rows after sprint twelve — every new sprint inherits the chart’s existing rows as context for the design decisions it has to make. The Build Spec Writer learned the company’s conventions in sprint one and applies them automatically in sprint two. The leadership team that was new to Work Deconstruction in sprint one runs it with judgment in sprint four. The training material from sprint three’s Deliver phase becomes the onboarding for sprint seven. None of this is a metaphor. These are real artifacts, sitting inside the company, accelerating the next sprint.

A member six sprints in finds the Source phase of their seventh sprint taking days instead of weeks. The constraint they are working on now is sharper, because three sprints’ worth of Sprint Retrospectives taught them what kinds of constraints actually move the headcount math. The Hybrid Accountability Chart entry they draft in Design takes an afternoon, because they have drafted six of them. That is the infrastructure compounding. It is not a speech the CEO gives at the offsite. It is what the work feels like in the seventh sprint, compared to the first.

Compound is not the same as scale.

Be precise here. The audience for this book has been pitched “scale your AI” in every form imaginable and has every right to be skeptical.

Scale is generic. It is the word every vendor uses when they want you to spend more money on the same thing. Scale tells you nothing about whether the second thing is cheaper than the first, or whether the second thing is better. It is a quantity claim with no architecture behind it.

Compound is specific. The infrastructure built by each sprint accelerates the next one — measurably, in artifacts you can point at. The cost per sprint decreases, because Source is faster, Design is faster, Build is sharper, and Deliver inherits training material from the previous sprints. The return per sprint increases, because the constraints get more strategic as the easy ones come off the backlog and the leadership team’s design judgment improves. The gap between revenue growth and headcount growth — the headcount math from the opening of this book, the dynamic everyone feels but nobody has named — starts to look like a solved problem. Not because revenue stops growing. Because revenue can grow without the head count line growing with it, and the company that runs the Rhythm long enough watches that gap widen on purpose.

That is the difference between compounding and scaling. Scaling is doing more of the same thing. Compounding is doing the same thing with infrastructure that makes the next iteration cheaper and better than the last.

The brand and the stage share a name on purpose.

By now you have noticed that the word Compound is doing more than one job in this book. It is the name of the company. It is the verb the work produces. It is the stage of the Sequence you are reading about now. The naming is load-bearing.

Compound the stage is where the verb happens — where the previous sprint’s output stops being an outcome and becomes the next sprint’s accelerant. That mechanism is what Compound the company exists to install. The brand and the stage share a name because the company is the stage — the institutional answer to “who teaches you to run this last step, and to run it every time?” The verb is what you are doing once the stage is part of how you operate.

Members get this on their second sprint. The first sprint feels like a project. The second sprint, when the Source map already exists and the chart already has a row and the design conversation has a precedent — that is when the word starts to mean what it means. Stop transforming. Start compounding. The tagline is not a slogan. It is a description of the operating posture the stage produces.

One sprint, one Compound.

One Sprint produces one Compound stage. That is the unit. You run the Sequence on one validated constraint, you ship the work, and you close it with Compound — Sprint Retrospective, Constraint Re-rank, one design change installed, the backlog re-ranked, the chart updated, the next constraint named. That is a complete pass through the system, and at the end of it your company knows things, and owns things, it did not at the start.

The chapter that follows is the Rhythm — what happens when this stage runs quarterly and becomes how the company operates. One Compound stage proves the mechanism works on one sprint. The Rhythm is what turns the mechanism into a business outcome. It is also where the headcount math from the opening of this book gets answered, in the language of a company that has been running the Sequence long enough that the answer is no longer hypothetical.