1. Product 2. Agents 3. Governance 4. CloudOps 5. FinOps 6. Security

AI + Data + Cloud · Pillar 1

📋

Product Management

Business-led Roadmap

6Golden Path Phases

Business-led roadmap for digital products. 6-phase golden path: Discover → Specify → Plan → Build → Measure → Iterate. 10 essential commands. 42 PM skills available in the marketplace.

⚡ First governed sprint in <4 hours

AI agents build governed & Humans ship trusted. 80% autonomy & 100% accountability.

Section One (Ch.1-7)

Creating the Transformation Roadmap

“When evaluating stalled digital and AI transformations, many of the issues can be traced back to insufficient planning and alignment. We typically see five serious sins: leaders have a different conceptual understanding of digital; leaders focus energies on pet projects that don't deliver much value; they focus too much on technology solutions.”

A business-led roadmap is the blueprint for a successful digital and AI transformation. Three elements are foundational and universal: vision, alignment, and commitment. ADLC enforces this through the product-owner agent and the Authority Chain: HITL -> Product-Owner -> Cloud-Architect -> Specialist — zero exceptions.

Platform Evolution

As Claude models improve reasoning, /speckit.specify generates higher-quality specs with fewer iterations. Each model release shortens the spec-to-code cycle.

Product Management Golden Path

Each phase answers: Who does it, Why it matters, What if you skip it

1Discover

Find problems worth solving. Validate assumptions before building.

lean-canvas + opportunity-solution-tree + /product:pr-faq

Who: HITL steers direction, product-owner agent validates business fit

Why: Building without discovery = building the wrong thing. Most digital transformations fail to capture expected value because they skip this step.

Skip? Pet projects, wasted sprints, no customer alignment

→ PR/FAQ + Validated Assumptions + INVEST Stories

2Specify

Turn validated ideas into precise specs. What to build, not how.

/speckit.specify + /speckit.clarify

Who: product-owner agent drafts, HITL approves scope

Why: Ambiguous specs = scope creep + rework. Spec-Driven Development makes specifications executable, not disposable.

Skip? Unbounded scope, 'just code it' syndrome, rework loops

→ spec.md with INVEST user stories + acceptance criteria

3Plan

Design the technical approach. Map agents to tasks. Constitution check.

/speckit.plan + /speckit.tasks

Who: cloud-architect agent designs, product-owner agent validates business alignment

Why: 80% of struggling transformations fail on scoping, not execution. Planning prevents over/under-engineering.

Skip? Over-engineering, no agent delegation, blocked dependencies

→ plan.md + tasks.md + agent delegation matrix

4Build

Execute tasks with specialist agents. PDCA until quality gates pass.

/speckit.implement + specialist agents

Who: AI agents build and execute, HITL reviews and ships trusted

Why: Agents prepare, humans decide. Governed execution with evidence prevents drift and testing theater.

Skip? Unreviewed code, no evidence, testing theater, NATO violations

→ Working software + test evidence in tmp/

5Measure

Track sprint health, DORA metrics, agent consensus scores.

/metrics:daily-standup + /metrics:sprint-review

Who: observability-engineer agent collects, HITL reads the dashboard

Why: What gets measured gets managed. Evidence-based ceremonies prevent 'done' claims without proof.

Skip? Invisible blockers, no sprint visibility, NATO violations

→ DORA dashboard + sprint evidence + ceremony artifacts

6Iterate

Retrospective, extract patterns, improve the process itself.

/metrics:sprint-retro + /speckit.retrospective

Who: Full agent team reflects, HITL decides what to change

Why: Digital transformation is a team sport. The process that builds products must itself improve continuously.

Skip? Same mistakes repeated, process stagnation, no learning

→ Improvement actions + updated constitution + PDCA evidence

Start Here

Spec-Driven workflow and product skills — copy/paste to start

.specify/ Workflow

6 sequential commands — run top to bottom to go from idea to sprint

1./product:pr-faq → North-star document (WHO + PROBLEM + METRIC)

2./speckit.specify → spec.md with INVEST stories + acceptance criteria

3./speckit.clarify → 5 targeted questions to remove ambiguity

4./speckit.plan → plan.md with architecture + cost estimate

5./speckit.tasks → tasks.md with dependency-ordered work items

6./speckit.implement → Working code + test evidence in tmp/

⚡ Idea to sprint plan in <4 hours

Product Skills & Ceremonies

Use alongside the workflow — discovery before /speckit.specify, ceremonies after /speckit.implement

1.lean-canvas → Business model hypothesis on one page

2.opportunity-solution-tree → Map outcomes to experiments

3.competitive-battlecard → Know your landscape before competing

4./metrics:daily-standup → Sprint health + DORA + recommendation

5./metrics:sprint-review → Demo items + completion metrics + DORA

6./metrics:sprint-retro → 4L retrospective + improvement actions

⚡ Full sprint cadence from day 1

Team & Governance

Link to agent team + live project oversight.

1./talent-bench

2./project?id=finops-focus

⚡ Cross-pillar discovery

Component Map

10 components implementing this pillar

Type	Name	Why	Business Value
Command	`/product:pr-faq`	Amazon-style PR/FAQ north-star document	Start here — define what you're building and why before any code
Skill	`lean-canvas`	Business model hypothesis on one page	Test before you build — riskiest assumptions first
Command	`/speckit.specify`	Create structured specs from natural language	Specs become executable — not disposable documents
Command	`/speckit.clarify`	5 targeted clarification questions	Surface ambiguity before it becomes rework
Command	`/speckit.plan + /speckit.tasks`	Implementation plan with dependency-ordered tasks	Design-before-code prevents scope creep
Agent	`product-owner (opus)`	Business validation FIRST before any specialist work	Authority chain: HITL → PO → CA → Specialist — zero exceptions
Command	`/speckit.implement`	Execute tasks with evidence collection	Agents build governed, humans ship trusted
Command	`/metrics:daily-standup`	Sprint health + DORA + single recommendation	Daily visibility — what gets measured gets managed
Command	`/metrics:sprint-review`	Demo items + completion metrics + DORA actuals	Evidence-based sprint closure
Command	`/metrics:sprint-retro`	4L review + DORA + improvement actions	Continuous process improvement across sprints

Risk & Scalability

What happens without this pillar, and why ADLC scales from 1 person to enterprise

What if you skip?

Without a business-led roadmap, teams build technology for technology's sake. Industry research finds only 31% of companies capture expected revenue lift from digital initiatives. The five serious sins: different conceptual understanding, pet projects, too much tech focus, spread too thin, CEO delegates away.

Scalability

The product-owner agent and speckit commands provide sprint governance that scales with your team. One manager with AI agents or a 100-person org — the ceremonies and evidence requirements remain consistent.