Agile and Scrum were built to protect one scarce resource: engineering hours. Building was slow and expensive, so teams shortened the loop and rationed the build. Sprints, standups, story points, the retro, the backlog grooming. Every ritual exists to keep hours from being spent on the wrong thing.
Agents removed that scarcity. A working implementation of a user story now takes minutes. The cost of building the wrong thing collapsed; the cost of not knowing what the right thing is went up, because the wrong thing now ships faster than anyone can catch it. A process designed to ration the build has nothing left to ration. In Rootshift, Agile and Scrum are deprecated. The discipline moves to the truth the software is derived from.
The vocabulary map
| Deprecated | In Rootshift |
|---|---|
| Scrum | Model iterations on Rootshift |
| Workshops | Knowledge intake |
| Backlog | Ground Truth |
| Tickets | Capabilities |
| Estimations | Data-backed simulations |
| Story points | Delivered history |
| Sprints | Delivery loops |
| QA | Evals and gates |
| Velocity | The record |
| Documentation | Owned, versioned artifacts |
Why each word changes
Scrum → model iterations on Rootshift. Each iteration updates the model of the business first: sources read, decisions recorded, artifacts approved. Agents then build from what was approved.
Backlog → Ground Truth. A backlog is a queue of requests. A Ground Truth is a model of what is true, with owners and approval state. Agents build from it without guessing.
Tickets → capabilities. A story is a unit of work. A capability is a unit of value: a business function, a verb, such as scan a receipt, reconcile a statement, invite a collaborator. It is made of stories with acceptance criteria and evals, and carries a class, a tier and a size so it is comparable across projects.
Story points and estimations → delivered history and data-backed simulations. Delivery is recorded run by run, so a forecast comes from what comparable capabilities actually took, not from a negotiation in a planning meeting.
Sprints → delivery loops. The loop is knowledge intake, Ground Truth, agent fleet execution, closed loops, quality gates, and telemetry. A change closes when its evidence is recorded.
QA → evals and gates. Every acceptance criterion has a named eval. Independent reviewer agents check the change, and a merge gate refuses work without evidence. No agent judges its own work.
Velocity → the record. Every run records outcomes, review findings found and held, decision latency, and bounces. That record makes versions comparable. People approve; agents build and verify.
Documentation → owned, versioned artifacts. Documentation finally has a reader that acts on it. Every claim needs an owner, a source, and a state.
What still holds
Certainty is earned by shipping. A validated model is enough to start. A mistake is cheap, because it is a Ground Truth fix and an agent re-run rather than weeks of rework. The ceremonies are gone; shipping to learn what is true is not.
What people do instead of ceremonies
People set the objectives, resolve the decisions the system queues, approve what is true, and accept what goes to production. The time that went into ceremonies goes into judgment. For the full argument, read the foundational essay.