What Mob Elaboration produces
In the AI-DLC Inception phase, the team sits with the AI on one screen. The AI proposes user stories, acceptance criteria, and units. The product owner, developers, and QA — the mob — correct what is under-built or over-built and line it up with real constraints.
A finished Mob Elaboration leaves user stories, non-functional requirements, risks, measurement criteria tied to the intent, and a suggested bolt. The method keeps user stories on purpose, as the contract for what gets built.
AI-DLC uses AI-driven for this shape of work: the AI proposes the breakdown, and people approve it. AI-assisted is the broader name for a person leading a task while the AI helps with a piece of it.
The same jobs, two vocabularies
AI-DLC renamed the familiar Agile objects so the shorter cycles have their own words. The jobs line up like this.
| AI-DLC | What it is | Scrum analog |
|---|---|---|
| Intent | The business or technical goal the work starts from. | Product goal, or a raw backlog idea |
| Mob Elaboration | The team and the AI turn that intent into stories, acceptance criteria, and units. | Backlog refinement |
| Unit | A cohesive slice that can be built on its own. | Epic |
| User stories | The contract between the people and the AI for what to build. | User stories |
| Bolt | A short build-and-check cycle, measured in hours or days. | Sprint |
| Mob Construction | The build session: domain model, design, code, and tests, with the team correcting the AI as it goes. | The sprint, run as mob programming |
Backlog health after the session
A useful Scrum refinement leaves about a sprint of ready work before planning: pointed, unblocked, and still current. Sprint Vitals reads that on the Jira board as backlog health — refined points against trailing velocity — plus blocked stories, stale tickets, and scope that changed after the team committed.
A useful Mob Elaboration leaves the intent as validated stories, acceptance criteria, units, risks, and a suggested bolt. That is the same readiness question under the AI-DLC names. The AI-DLC method itself asks whether story points and velocity still describe the work once an AI flattens easy and hard tasks, and it points teams toward business value as well.
Sprint Vitals reads a Jira Scrum backlog. When the team still keeps that work in Jira, the backlog health snapshot is the check after refinement. Bolts and units live in the AI-DLC artifacts, so the Jira number describes the session only while the ready work is still on the board.
Nearby methods use other names for the same moment. BMAD calls the multi-agent room party mode, and it prepares a story with story review before an implementation-readiness check. Spec-driven development calls the pass clarify, then specify. Older Scrum teams still say grooming for refinement.
Where this shows up in Sprint Vitals
The definition of ready is the agreement a Scrum team uses in refinement. Sprint Vitals puts backlog health on the board for the next three future sprints, so the team can see whether refinement covered about a sprint of ready work before planning.
A Jira admin approves Marketplace apps for the site. If you are the scrum master, send them the install link. If you already have Sprint Vitals, the manual explains how to read those numbers on the board.