Agile Frameworks - SAFe
About The Author & The Article
Jonathan Bishop, Group Chairman, Bishop Phillips Consulting. [1]
Copyright 2020-2026 - Moral Rights Retained.
This article may be copied and reprinted in whole or in part, provided that the original author and Bishop Phillips Consulting is credited and this copyright notice is included and visible, and that a reference to this web site (http://RiskWiki.bishopphillips.com/) is included.
This article is provided to the community as a service by Bishop Phillips Consulting www.bishopphillips.com.
Introduction
SAFe (Scaled Agile Framework) is a way to align strategy, funding, architecture, and multiple Agile teams using a predictable cadence and multi‑team coordination mechanisms so the entire enterprise delivers value continuously and coherently.
It’s Agile for organisations with:
- multiple teams
- shared platforms
- shared architecture
- regulatory constraints
- long‑term funding
- enterprise‑level governance
Scrum is for teams, while SAFe is for systems.
The Four Levels of SAFe (the real structure)
SAFe has four layers which are effectively 'zoom levels' in a value‑delivery system.
1. Team Level (Scrum/Kanban Teams)
This is the level we know from Scrum.
Each team uses:
- Scrum or Kanban
- PBIs
- Sprint Goals
- Definition of Done (DoD)
- Retrospectives
Nothing changes from Scrum at this level. SAFe doesn’t replace Scrum: it connects Scrum teams.
2. Program Level (Agile Release Train — ART)
This level is the 'heart' of SAFe.
An **ART** is:
- 5–12 Agile teams
- working on one product or value stream
- synchronised on the same cadence
- delivering together every 8–12 weeks
An ART as a Scrum of Scrums with structure.
Key concepts:
- PI (Program Increment) = 5 Sprints
- PI Planning = all teams plan together
- System Demo = integrated demo across teams
- RTE (Release Train Engineer) = the Scrum Master for the whole train
- Product Management = the Scrum Product Owner at scale
- System Architect = architecture at scale
This solves the “multiple teams, one product” problem.
3. Large Solution Level (for big systems)
Used when multiple ARTs must coordinate to build a single large system.
Examples:
- Defence systems
- Banking platforms
- National infrastructure
- Enterprise‑wide platforms
Roles:
- Solution Train Engineer
- Solution Architect
- Solution Management
If you don’t have massive systems, you can ignore this layer.
4. Portfolio Level (Strategy + Funding)
This is where SAFe becomes Enterprise Agile.
Portfolio level handles:
- Lean budgeting
- Value streams
- Epic approval
- Investment horizons
- Enterprise architecture
- Governance
Key artifacts:
- Portfolio Kanban
- Strategic Themes
- Epic Owners
- Guardrails
This is where SAFe connects strategy to execution.
The Core SAFe Idea: Value Streams
SAFe organises the enterprise around value streams, not departments.
A value stream is: The sequence of activities that delivers value to a customer.
Each value stream has:
- funding
- teams
- architecture
- governance
- metrics
This is the opposite of traditional org charts.
The SAFe Cadence
SAFe runs on a predictable rhythm:
- Daily: Team standups
- Every 2 weeks: Sprints
- Every 10 weeks: PI (Program Increment)
- Every PI:
- PI Planning
- System Demo
- Inspect & Adapt
This creates alignment across the entire enterprise.
The SAFe Roles
SAFe has three major layers of roles:
- Team Level — where work is done
- Program Level (ART) — where multiple teams align
- Portfolio Level — where strategy and funding live
There is also a Large Solution Level for massive systems.
These three levels have the following roles:
1. Team Level
This is the Scrum level which are explained in detail in Agile Frameworks - Scrum:
- Product Owner
- Scrum Master
- Developers
2. Program Level
This is where SAFe becomes SAFe. The ART is 5–12 teams working together on one product/value stream:
- Product Manager
- Release Train Engineer (RTE)
- System Architect
Plus two external roles:
- Business Owners
- System Team
| Role | Description |
|---|---|
| Product Manager | Owns the Program Backlog.
Focus: What should the ART build next?
PO = team-level value. PM = program-level value. |
| Release Train Engineer (RTE) | The Scrum Master for the entire ART.
Focus: How do all teams work together?
This is one of the most critical roles in SAFe. |
| System Architect / Solution Architect | Owns the architecture runway.
Focus: What technical direction ensures long-term viability?
|
| Business Owners | The key stakeholders for the ART.
Focus: Is the ART delivering business value?
|
| System Team | A specialised team that supports integration and testing.
Focus: Enable continuous integration and system demos. |
3. Portfolio Level
This is where SAFe becomes Enterprise Agile:
- Epic Owners
- Enterprise Architect
- Lean Portfolio Manager
| Role | Description |
|---|---|
| Epic Owners | Own portfolio epics (big initiatives).
Focus: Drive large-scale change.
Coordinate across ARTs |
| Enterprise Architect | Owns enterprise-wide architecture.
Focus: Ensure technical coherence across value streams.
|
| Lean Portfolio Manager | A group, not a person.
Focus: Where should we invest?
|
4. Large Solution Level (Optional)
This is where SAFe becomes Enterprise Agile:
- Solution Train Engineer (STE) - RTE for multiple ARTs.
- Solution Architect - Architect for the entire solution.
- Solution Management - Product Management at the solution level.
If you’re not building aircraft, banking platforms, or defence systems, you can ignore this layer.
These roles ensure alignment from strategy → execution.
The Real Relationships
PO vs PM
- PO = team backlog
- PM = program backlog
- PM tells PO what matters
PO tells team what to build next
Scrum Master vs RTE
- Scrum Master = team flow
- RTE = ART flow
- RTE coaches Scrum Masters
- Scrum Masters coach teams
Architects
- System Architect = ART architecture
- Enterprise Architect = portfolio architecture
- Solution Architect = multi-ART architecture
Business Owners
- They are the customers of the ART
- They approve PI Objectives
- They ensure the ART is delivering value
SAFe PI Planning
What SAFe PI Planning Actually Is
PI Planning (Program Increment Planning) is a two‑day, all‑hands planning event where every team in an Agile Release Train (ART) aligns on a shared mission, identifies dependencies, resolves risks, and commits to a set of objectives for the next Program Increment. It is a big-room planning event where all teams in an ART, usually 50–125 people, come together for two days to:
- align on a shared mission
- agree on a set of objectives
- identify dependencies
- surface risks
- commit to a plan for the next 8–12 weeks
It is the heartbeat of SAFe. If the ART is the engine, PI Planning is the ignition cycle.
Why PI Planning Exists
Scrum works beautifully for a single team. But when you have 10 teams working on one product, you get:
- conflicting priorities
- hidden dependencies
- integration failures
- architectural drift
- duplicated work
- misaligned goals
PI Planning solves this by creating:
- shared context
- shared priorities
- shared cadence
- shared commitment
It is the moment the entire ART synchronises.
The Structure of PI Planning (2 Days)
DAY 1 - Alignment + Draft Planning
1. Business Context (Leadership)
Executives explain:
- market conditions
- customer needs
- strategic priorities
- upcoming deadlines
- financial constraints
This sets the "why."
2. Product Vision & Roadmap (Product Management)
Product Managers present:
- vision
- top features
- priorities
- architectural runway
This sets the "what."
3. Architecture Vision (System Architect)
Architects explain:
- technical direction
- constraints
- enablers
- risks
- integration concerns
This sets the "how."
4. Team Breakouts - Draft Plans
Teams break out and:
- pull features from the Program Backlog
- break them into stories
- estimate
- identify dependencies
- raise risks
- draft PI Objectives
This describes the "what".
5. Draft Plan Review
Teams present their draft plans to:
- Product Management
- RTE
- Business Owners
- Other teams
Feedback is given, dependencies are negotiated and scope is adjusted.
DAY 2 — Finalisation + Commitment
6. Planning Adjustments
Teams refine their plans based on feedback.
7. Management Review & Problem Solving
Leadership meets privately to:
- resolve resource conflicts
- adjust scope
- address cross-team issues
- make trade-offs
This is where the “big rocks” get moved.
8. Final Team Breakouts
Teams update:
- stories
- dependencies
- risks
- PI Objectives
9. Final Plan Review
Each team presents:
- their plan
- their risks
- their dependencies
- their PI Objectives
Business Owners score each team’s objectives (1–10) based on business value.
10. ROAMing Risks
All risks are categorised as:
- Resolved
- Owned
- Accepted
- Mitigated
This creates transparency.
11. Confidence Vote
Everyone votes (1–5 fingers) on confidence in the plan.
If confidence is low: rework.
12. PI Planning Ends → PI Execution Begins
Teams start Iteration 1 the following Monday.
The Outputs of PI Planning
By the end, the ART has:
1. Committed PI Objectives
Each team has 5–10 objectives:
- some committed
- some stretch
2. Program Board
A visual map showing:
- features
- teams
- dependencies
- milestones
This is the ART’s “single source of truth.”
3. Identified Risks
All major risks are ROAMed.
4. Shared Understanding
Everyone knows:
- what we’re building
- why we’re building it
- how we’ll work together
This is the real value.
The Intent of the PI Planning Approach
It creates alignment at scale.
- Teams align with each other
- Teams align with Product Management
- Product aligns with Architecture
- Architecture aligns with Strategy
- Everyone aligns with the Business Owners
This is impossible to achieve through documents or asynchronous communication.
PI Planning is the social synchronisation mechanism of SAFe.
Common Misconceptions
"PI Planning is waterfall."
No — it’s *collaborative planning*, not *predictive planning*.
"PI Planning locks scope."
No - scope is flexible. Objectives are the commitment, not the stories.
"PI Planning is too big."
It’s cheaper than:
- misalignment
- rework
- integration failures
- architectural drift
Why SAFe Exists
SAFe solves problems that Scrum alone cannot:
- Multiple teams working on one product
- Shared architecture
- Shared platforms
- Regulatory compliance
- Long‑term funding
- Enterprise governance
- Cross‑team dependencies
- Coordinated releases
Scrum is too small for enterprise wide application so SAFe provides a scaffolding around Scrum but in the process it looses some of its agility. SAFe’s Secret: It’s Not Really About Agile. SAFe is actually about:
- alignment
- flow
- governance
- architecture
- funding
- strategy execution
Agile is the delivery mechanism while SAFe is the operating model at enterprise scale.
SAFe + Agentic AI (our future direction)
This is where things get exciting.
SAFe provides:
- structure
- cadence
- governance
- alignment
Agentic AI provides:
- automation
- analysis
- orchestration
- continuous sensing
- autonomous execution
Together, they create: Enterprise Agile 2.0: A hybrid human–AI operating model.
Agents can plug into:
- Portfolio Kanban (epic analysis)
- PI Planning (capacity simulation)
- Backlog refinement (PBI generation)
- Architecture runway (design evaluation)
- Flow metrics (bottleneck detection)
- Compliance (continuous monitoring)
This is the future we are moving towards.
Next Steps: Choose Your Path
- SAFe + Agentic AI (the future model) - How agents plug into every SAFe layer.
- SAFe Roles Explained - PO vs PM vs RTE vs Epic Owner.
- SAFe Portfolio Kanban - How strategy flows into execution.
- SAFe PI Planning - How multi‑team planning actually works.
- SAFe vs Scrum vs Kanban - When to use which.
- How to implement SAFe in a real organisation - Practical, not theoretical.
