ADOM / Process Explorer
Working model · v02
From product intent to delivery

Make the handoffs visible.

Explore how a request becomes an initiative, a delivery issue, and a sprint plan. Separate what we’ve observed from what we’re proposing.

Proposed process · from intake to sprint selection

Follow one request through the handoffs.

Choose a route, then follow the connected steps. Each step shows who prepares the work, what the review decides, and what moves forward.

Group structure and board visibility

Current model: SEC, Dispatcher Phoenix, and Dispatcher Stratus are separate group roots. SEC / Epic / Intake holds incoming issues centrally; delivery issues move to their owning product projects. SEC epics can coordinate those issues, but a SEC group issue board does not combine product-project issues across these roots.

A common ancestor is a separate option: shared issue views could span the product projects, letting known-owner requests stay in their owning project. This exploration uses the current separate-group structure. Board names and issue preparation stages below are proposed views that still require team agreement.

01 · Understand the levels

Keep the goal connected to the work.

Initiatives define the outcome. Capability epics define the capabilities. Issues define the specific changes teams deliver.

Proposed model · illustrative Fleet Analytics examples
01 · Initiative

The outcome

What result are we pursuing, and why?

A coordinated effort to achieve a customer or business outcome. It defines the shared goal, scope, accountable owner, and success measures.

Illustrative example

Initiative — Fleet Analytics and Reporting

Help administrators make better-informed fleet decisions. The specific outcome and measures still need agreement.

GitLab: Epic · level::initiative · SEC initiative view
Achieved through · explicit links
02 · Capability epic

The capability

What will users be able to do?

A substantial, bounded capability that contributes to the initiative. It defines functionality and clear completion criteria.

Illustrative example

Historical Analytics

Enable administrators to examine device usage over time. Other proposed capabilities include Standard Fleet Reports, Device Lifecycle Management, and Custom Report Builder.

GitLab: Epic · level::capability · linked to the initiative
Delivered through
03 · Delivery issue

The specific change

What will the team build and test?

A testable change with expected behavior, acceptance criteria, and a scope the team can estimate and deliver.

Illustrative example

Display usage changes over a selected period

One delivery issue contributing to Historical Analytics. Recording periodic device readings could be another.

GitLab: Issue · child item of the capability epic

“Initiative” describes a planning purpose in this proposal; it is represented by an Epic work item. Capability epics also use the Epic type. Issue is the delivery item type; “child item” describes its relationship to the epic. Standalone issues do not require an epic or initiative.

02 · Why the distinction matters

Breaking down the work preserves the goal.

Direction

Every capability and delivery issue connects to a customer or business objective.

Accountability

The overall outcome retains an owner after breakdown, across teams and releases.

Visibility

Product can review outcomes, capability delivery, and sprint work separately.

Completed issues contribute to a capability. Delivered capabilities contribute to an outcome. Neither automatically proves the next level is complete. Ownership, success measures, and regular reviews make the structure useful.

03 · Compare the change

Current practice and the proposed model

Observed practices and open questions, compared at the same planning levels.

01

Initiative

Current · observed

Original epic retained during breakdown

Fleet Analytics and Reporting is retained during breakdown and already has Linked items relationships to the resulting capability epics. Related titles also use a shared name prefix.

Working assumption
After breakdown, the overall initiative and its outcome are not tracked at a higher level; tracking continues through the resulting capability epics.

Proposed

Retain a clearly identified initiative

Use the title Initiative — Fleet Analytics and Reporting and level::initiative. Move it from intake into a dedicated initiative view, retaining an owner, overall status, success measures, capability summary, and review cadence.

02

Capability

Current · observed

Capability epics already linked

The resulting epics represent pieces of the larger initiative and are already connected through Linked items. A shared title prefix also signals their association; the links establish a relationship rather than native containment.

Proposed

Keep the links; clarify each epic’s purpose

Keep the existing Linked items relationships and apply level::capability to delivery capabilities. Each has its own scope and completion criteria. The initiative owner maintains their combined status and contribution to the overall outcome.

03

Sprint delivery

Current · observed

Epic structure reflects sprint scope

Teams choose child issues from epics labeled for the sprint. To represent the actual commitment, an epic may be divided into sprint-sized work containers, or given additional sprint labels as its issues are scheduled across sprints.

Effect
The capability breakdown and sprint plan become intertwined: epic structure or labels must change to reflect when its issues will be delivered.

Proposed

Separate capability scope from sprint allocation

Keep capability epics organized around the functionality being delivered. Select specific issues across epics and standalone work, check their combined effort against team capacity, and assign sprint iterations to those issues.

Benefit
A capability’s issues can span several sprints without dividing its epic into sprint containers. Track selected issues by team and sprint; keep the epic focused on capability progress and forecast.

Guiding recommendation

Leaving intake should change the initiative’s planning view, not end its accountability.

04 · How we manage it in GitLab

Put section 03 into practice: keep the capability structure stable, record sprint allocation on delivery issues, and review the overall outcome at initiative level. This is a proposed configuration, not the installed workflow.

1. Create and connect the working records

Record / locationConfigure in GitLabMaintain on the record
Initiative
Epic in SEC
Retain and rename the original epic Initiative — Fleet Analytics and Reporting. Apply level::initiative. Maintain a capability summary table in its description.Accountable owner, objective, scope, success measures, overall status, decisions, and review cadence.
Capability
Epic in SEC
Retain existing capability epics and their Linked items relationships; create additional capabilities only where agreed scope requires them. Apply level::capability.Capability owner, scope, completion criteria, dependencies, target dates, and delivery forecast.
Delivery issue
Owning Phoenix or Stratus component project
Create or retain the issue in its project. Add it as a child item of the capability epic. Standalone issues remain without an epic parent.Acceptance criteria, readiness, estimate / weight, assignee, proposed team label, and iteration when selected.

Use the confirmed Linked items → relates to relationship. Add it once; it appears on both epics. It does not identify which epic is the initiative, so the naming convention and level labels establish that meaning. Maintain initiative summaries manually; capability-to-issue relationships remain native child items.

Maintain the agreed record in the description: use a table with capability link, owner, status, target, and dependencies. Add a last-reviewed date. Comments support discussion; incorporate agreed decisions into the description. The initiative owner updates the summary at the agreed review cadence; linked epics do not update these fields automatically.

Naming and tracking policy

Initiative title: Initiative — [Name], with level::initiative. Capability label: level::capability. Use descriptive capability titles; a name prefix is optional for readability.

Initiative and capability are planning roles of the Epic work-item type. Keep initiative epics active after intake and breakdown, in a dedicated initiative board or filtered view. Assign an accountable owner and review cadence; complete or cancel the initiative after evaluating its outcome. Removing it from intake must not remove it from ongoing tracking.

2. Set up views for different decisions

SEC initiative view

Filter epics by level::initiative. When intake ends, keep the initiative active here with an accountable owner. Product reviews overall status, outcomes, risks, decisions, dependencies, and the next review date.

SEC capability view

Filter by level::capability. Review sequencing, targets, dependencies, child-issue progress, and forecasts across sprints.

Product issue views

Within each product group, use a preparation / ordered backlog view and a sprint issue board. Include epic children and standalone issues together.

Agree the issue-readiness labels and their owners before configuring the backlog. For sprint views, use an agreed team label (for example team::A) and the selected iteration. These are views of the same issues, not copies or additional epic containers.

Current group constraint: SEC, Phoenix, and Stratus are separate roots. A SEC group issue board does not combine their product projects. Use product-group issue views and link them from SEC for now; a common ancestor is a separate structural option. Team labels and iteration dates need coordination across the separate groups.

3. Select and record sprint scope on issues

  1. Product and the delivery team review ready issues across capabilities and standalone work, considering priority and dependencies.
  2. Check the selected issues’ combined estimates against available development and testing capacity. An iteration records the timebox; it does not decide capacity.
  3. Set the agreed team label, assignee, and sprint iteration on each selected issue. Use iterations available to that issue’s product group; leave unselected issues in the backlog.
  4. Filter the team’s issue board to that iteration. Track individual issue development, testing, and completion using the agreed workflow.

Example: Historical Analytics has ten issues. Assign three to Sprint 17 and others to later iterations when selected. Keep one Historical Analytics capability epic; do not divide it solely to match sprint boundaries.

4. Keep delivery and outcome reviews connected

During the sprint: update the selected issues and record blockers or scope changes. If work moves to another sprint, update its iteration and reassess the plan rather than reshaping its capability epic.

At capability review: inspect child-issue progress, dependencies, and remaining work; update target dates and the forecast. If SEC retains epic sprint labels, treat them as a maintained forecast summary, not the record of exact sprint scope. Do not assume they synchronize automatically with child iterations.

At initiative review: the initiative owner consolidates capability status and evaluates success measures. Maintain the capability table and an overall narrative status manually. Keep the initiative active until its outcome is delivered, reduced, or cancelled; completed issues alone do not establish success.

5. Native nesting: optional improvement

Check the installed version, license, and permissions. Ordinary epics require Premium or Ultimate; nested epics require Ultimate. Verify adding existing product-project issues beneath SEC capability epics; cross-group links do not expand board scope.

If native nesting is confirmed available later, connect capability epics as children of the initiative epic and use supported progress summaries. Retain outcome reviews and the accountable owner. Until then, explicit links and the manually maintained summary are the implementation plan; issue-level sprint allocation works independently of epic nesting.

GitLab references: Epics and hierarchy ↗ · Issue boards ↗ · Iterations ↗. Feature availability must be checked against the installed version.

Illustrative planning sandbox

Forecast the epic. Select the issues.

Choose from a mixed backlog. Points below are invented for demonstration, not estimates of KM work. Capacity is a discussion input, not a guarantee.

Selected effort
0 / 16
Selected issues
0
Plan check
Select ready work

Ordered candidate backlog

Sprint selection

Team A · illustrative next sprint

Not ready / deferred

Production Print device authentication

Epic child · outside November scope · unestimated

Not eligible yet
Evaluate alternative OCR engines

Discovery request · accuracy requirements unresolved

Investigate

Agree on the operating rules.

Use these questions to test the model with one epic child and one standalone request. Checking a box records your review locally; it does not indicate organizational approval.

Saved only in this browser. Use export to bring notes back into our next iteration.

Research opportunity · SEC availability unverified

GitLab Duo — AI-assisted planning

Help prepare and analyze the work. Keep people accountable for the decisions.

What is GitLab Duo?

GitLab’s integrated AI platform supports planning, development, testing, and delivery through conversational assistance and specialized agents working with available GitLab context.

Why explore the Planner Agent? It is the planning-focused agent most relevant to this proposal: potential assistance with work definition, breakdown, backlog analysis, and status reporting.

This explorer demonstrates possible uses. It does not connect to GitLab Duo or run these prompts.

The process still comes first

Duo assists the process; it does not define it.

SEC still needs agreed initiative tracking, ownership, relationships, requirements documentation, prioritization, and readiness rules. Drafts and AI-generated estimates require human review before becoming agreed delivery requirements.

01 · Where assistance could fit

The same ADOM steps, with help preparing them.

Proposed SEC applications, not confirmed functionality on the SEC instance.

ADOM activityPotential Duo assistanceHuman decision / review
Product intakeSummarize requests; flag missing information and possible duplicates.Confirm the problem, evidence, duplicates, and route.
Epic definitionDraft objectives, requirements, scope, and acceptance criteria.Agree scope, owner, outcome measures, and exclusions.
Epic breakdownPropose capability areas, user stories, and delivery issues.Validate useful slices, feasibility, boundaries, and testability.
Technical reviewSurface documented dependencies, blocked work, and possible gaps.Engineering / QA validate dependencies, security, and technical conclusions.
Backlog refinementFind stale items, unclear requirements, and incomplete records.Decide what to clarify, reprioritize, defer, or close.
Sprint planningSummarize candidates, suggest relative effort, and surface risks.The team validates estimates, dependencies, capacity, and exact issue selection.
Progress reportingSummarize completion, remaining work, and blockers.Verify source records and separately assess initiative outcomes.
02 · Try the idea with Fleet Analytics

Prompt examples for an initiative and its capabilities

Supply the initiative and relevant capability / issue URLs. Fleet Analytics uses related epics, so explicitly name that structure rather than assuming the agent will traverse a native child-epic hierarchy.

Examples are draft evaluation prompts. Results depend on accessible context, installed capabilities, and permissions. Any claims of missing work or dependencies need verification.

03 · Availability and practical limits

GitLab documents Planner Agent support for self-managed installations. SEC access is unverified. Its documented history begins with beta in 18.6, create/edit support in 18.7, and general availability in 18.8; newer actions vary by version. These facts do not establish what SEC has enabled.

  • Administration: confirm installed version, Duo entitlement / credits, cost, and foundational-agent configuration.
  • Actions: confirm whether analysis, creation, editing, and linking are supported and permitted. Duo access does not establish availability of native nested epics.
  • Context: test coverage across SEC, Phoenix, and Stratus; give explicit URLs and check whether all related items were read.
  • Governance: confirm security, data access, AI-use approval, and administrative requirements before using internal records.

Documented limitations: large item collections can reduce response quality; long comment histories may be incomplete; verify reported edits on the records themselves. Keeping agreed requirements current in descriptions improves the evaluation’s source material.

04 · Suggested limited evaluation

Start with preparation and summaries.

  1. Confirm access and governance with the appropriate SEC administrators.
  2. Use one existing capability epic and a small set of its issues. Request a requirements review and draft breakdown without editing records.
  3. Compare the result with Product, Engineering, and QA review. Measure preparation time, correction effort, coverage, and usefulness.
  4. Test an initiative summary using explicit links. Check whether completed work, blockers, and remaining scope match the source records.

Recommendation: evaluate whether Duo reduces manual preparation and improves issue clarity and summaries. Expand only after the team judges the results useful and agrees any permitted write actions.

Status: research opportunity — not an approved SEC capability or implementation plan.

Official references · reviewed November 4, 2026

GitLab Duo overview ↗ · Planner Agent documentation ↗ · GitLab Duo agents ↗

The feature summaries describe current documentation, not verified SEC availability. Mapping the capabilities to ADOM is a proposal for evaluation.