Open-source operating system for product teams: integrate AI into workflows without losing human accountability.
Keywords: product management · AI workflow · operating model · framework · methodology · human-in-the-loop · product operations · governance · enterprise AI
Your team adopted AI tools in the last 12 months. Work moves faster. So why does the brief still arrive incomplete at engineering? Why does scope shift mid-sprint? Why has no one reviewed the metrics 30 days after the last release?
AI accelerates the shuttle. Without a warp — stable structure, clear handoffs, named owners — it accelerates the tangle, not the fabric.
ORDITO is the warp.
If you can do only one thing: change how you write tickets.
Every Feature Request needs three things before it moves forward:
- A specific problem — not "users want X" but "enterprise admins spend 3h/week on Y because Z"
- A measurable outcome — not "better UX" but "reduce onboarding time below 2 days"
- At least 3 acceptance criteria — in given/when/then format
That's 20% of ORDITO. It produces 80% of the value. Start there.
When that feels natural, add the next layer. The 30/60/90 plan shows you exactly how.
The name comes from the Italian word for warp — the longitudinal threads held under tension in a loom while the weft passes through them to create fabric. Without a tense, regular warp, every pass of the weft produces tangle instead of pattern.
In a product team, the warp is principles, gates, contractual artifacts, roles, metrics — the stable structure. The weft is initiatives, features, decisions — the variable work that passes through it. AI accelerates the shuttle, but only if the warp is well tensioned.
ORDITO solves a concrete problem: teams that adopt AI tools without an underlying structure produce chaos faster. The framework provides the structure — gates, contractual artifacts, clear roles, learning loops — so that AI accelerates work instead of amplifying noise.
ORDITO does not replace Scrum, SAFe, or Shape Up. It lives on top of them, focusing on handoffs between roles and on native AI integration in the product lifecycle.
| If you are... | ORDITO helps you... |
|---|---|
| Product Lead overwhelmed by fragmented requests | Filter intake, prioritize with explicit criteria |
| Engineering Lead tired of incomplete briefs | Get handoffs with AC, tracking, and risk already structured |
| UX Lead dealing with handoffs that get lost | Trace every design decision through to release |
| Sponsor without portfolio visibility | See the status and expected value of every initiative |
| Team that just adopted an LLM | Know where AI accelerates and where it does damage |
┌─────────────────────────────────────────────────┐
│ PLAYBOOK — variants by context │ ← Adaptation
├─────────────────────────────────────────────────┤
│ OPERATING — roles, artifacts, rituals │ ← Application
│ MODEL AI services, tool integrations │
├─────────────────────────────────────────────────┤
│ FRAMEWORK — principles, gates, metrics │ ← Stability
│ human override policy │ (the warp)
└─────────────────────────────────────────────────┘
ORDITO recognizes that the same team does different things and adjusts its tension:
- ORDITO Core — existing product, stable team. Standard discovery, full gates, traceability.
- ORDITO Explore — new MVP, uncertain market. Exploration sprint, kill criteria, narrow rollout.
- ORDITO Scale — complex legacy, multi-team, compliance. Dependency mapping, architecture sign-off, audit trail.
- The Manifesto — the 10 principles that hold the framework together
- The three levels — how it's structured
- End-to-end example — a real feature from intake to learning
Follow the 30/60/90 plan. Summary:
- Days 0-30 — Stabilize intake and discovery (artifact templates + 2 AI services)
- Days 31-60 — Connect design, engineering, and QA (G2 + code review agent)
- Days 61-90 — Close release and learning loop (release checklist + impact review)
Want to contribute? Read CONTRIBUTING.md
Evolution proposals go through RFCs (the same model the framework proposes — consistency above all).
ordito/
├── assets/brand/ # Logo and social preview assets
├── 01-framework/ # Principles, gates, metrics, regulatory mapping
├── 02-operating-model/ # Roles, modes, rituals, tension modes
├── 03-artifacts/ # Schemas, templates, compiled examples
├── 04-ai-service-mesh/ # 13 AI services with prompts, evals, governance, profiles
├── 05-playbooks/ # Variants for Core / Explore / Scale / Hotfix
├── 06-workflows/ # Mermaid diagrams of processes and swimlanes
├── 07-adoption/ # Onboarding plan, antipatterns, change management
├── 08-integrations/ # Tool-specific integration guides + methodology mappings
├── schemas/ # Machine-readable JSON Schemas
└── tools/ordito-cli/ # CLI: scaffold, validate, gate, export
- Logo — used in this README header
- Social preview card — intended for repository social previews and link unfurls
- Machine-readable artifacts. Every artifact has a JSON Schema. CI validates that every template conforms. AI services read and produce the same contract.
- Tracked overrides. AI can be ignored, but the pattern of overrides is the system's tuning signal.
- Privacy by design. Every AI service declares its model, retention, and output visibility. Ready for enterprise environments.
- Three real modes. Not one imposed size. Core, Explore, and Scale have different documentation tension.
- Diff-able diagrams. Workflows in Mermaid, not PNG. They live in the repo as text, reviewable in PRs.
- v1.1 — Manifesto, schema, swimlane Core, end-to-end example
- v1.2 — Principles, roles, rituals, 30/60/90 plan, unified swimlanes, 12 AI services (registry + governance), all examples
- v1.3 — Prompt templates for 7 core AI services + Jira integration guide
- v1.4 — Remaining 5 prompt templates + adversarial
consistency-checker; antipattern catalog; schema v2 (lifecycle/gate split, mode, autonomy, blast_radius, prompt_version); methodology mappings (Scrum/SAFe/Shape Up/Kanban); tension modes (Overlay/Standard/Strict); CLI skeleton; eval framework; agent profiles; EU AI Act + NIST AI RMF mapping; Principle 7 reframed (substitution vs augmentation) - v1.5 — CLI v0.2 (
gatewrites back, offlineevalreadiness runner); Linear and Notion integration; Jira ↔ JSON sync; reference harness for AI services - v1.6 — Case studies from adopting teams; production-grown golden sets per service
- v1.7 — Audit-pack exporter (CLI v0.4); deeper EU AI Act high-risk QMS templates; sub-processor / DPA registry fields
Apache 2.0 — see LICENSE. You can use, modify, distribute, and adopt ORDITO in commercial contexts. We only ask for attribution.
The current manifesto, metaphor, and structure are the result of iteration with real teams in enterprise contexts.
- Discussions → GitHub Discussions
- Issues → GitHub Issues
"The warp is the threads that stay still. The weft is what passes through them. Without a warp, the weft is just a tangle of yarn."