SAFe SSM cheat sheet

Scrum is an empirical process-control framework resting on three pillars: transparency, inspection, and adaptation.

The five Scrum values are commitment, focus, openness, respect, and courage.

Scrum defines three accountabilities - Product Owner, Scrum Master, and Developers; SAFe adds the RTE (ART level) and STE (Solution Train level).

Scrum artifacts are the Product Backlog, Sprint Backlog, and Increment; each carries a commitment - Product Goal, Sprint Goal, and Definition of Done.

Scrum events are the Sprint (the container), Sprint Planning, Daily Scrum, Sprint Review, and Sprint Retrospective.

In SAFe the Sprint is called an Iteration, and Iteration Planning replaces Sprint Planning as the term used across the ART.

SAFe extends Scrum with XP engineering practices (CI, TDD, refactoring, pair work) and Kanban to form the Team and Technical Agility competency.

SAFe adds ART-level events - PI Planning, System Demo, ART Sync, and Inspect & Adapt - because Scrum alone does not coordinate multiple teams.

Scrum has no project manager or team lead role; cross-team coordination is handled by ART events and by the RTE, not by a command hierarchy.

The SAFe Scrum Master is a servant leader and coach for one Agile Team (sometimes two), with no people-management authority over team members.

Core responsibilities are coaching the team in Scrum and SAFe, facilitating team events, removing impediments, and fostering self-management.

The SM owns the process - teaching and upholding Scrum rules - while the team owns its technical and delivery decisions.

Impediments within team authority are resolved by the SM; impediments beyond it are escalated to the RTE or STE.

The SM helps the team build its Kanban board, set WIP limits, and use flow metrics instead of assigning tasks.

The SM supports the Product Owner by facilitating backlog refinement, story splitting, and definition of PI objectives, without making product decisions.

Coaching progression follows Teaching → Coaching → Mentoring → Facilitating as the team matures toward self-management.

The SM protects team time and capacity by filtering interruptions and aligning with the RTE on ART-level commitments.

SMs participate in the Scrum Master/Team Coach Community of Practice to share patterns and improve ART-level flow.

PI Planning is a two-day, cadence-based event at the start of each Program Increment where all ART teams align on a shared mission and draft plans.

Standard agenda - Day 1: business context, product/solution vision, architecture and development practices, team breakouts; Day 2: draft plan review, management review and problem solving, final plan, confidence vote.

Each team writes PI Objectives that are SMART, and Business Owners assign each objective a business value on a 1–10 scale.

Committed objectives are what the team will deliver; uncommitted (stretch) objectives are aspirational and excluded from predictability.

Risks are ROAMed: Resolved, Owned, Accepted, or Mitigated, each with a named owner.

Dependencies are recorded on the ART planning board (program board) with team-to-team links mapped to iteration timing.

The confidence vote is a fist-of-five; a team average of 3 or more allows the plan to proceed, and lower scores trigger replanning.

Scrum Master pre-work: confirm team capacity, ensure backlog readiness, secure logistics, and prepare the team for breakouts.

During breakouts the SM facilitates and timeboxes, keeps the team focused, surfaces dependencies, and captures the team's draft plan.

Iteration Planning (typically two hours or less for a two-week iteration) produces the Iteration Goal, selected stories, and a task breakdown against capacity.

Capacity is baselined on prior velocity, with allowance for ART-level work and unplanned items.

The Daily Stand-up is a 15-minute, same-time, same-place event focused on progress toward the Iteration Goal and on impediments - not a status report to the SM.

The team Kanban visualizes flow; explicit WIP limits and explicit policies expose bottlenecks and enforce pull-based work.

The Iteration Review is a team-level demo of the working increment to stakeholders, plus a metrics review and backlog adjustment.

The Iteration Retrospective inspects process and team dynamics and yields a small number of improvement actions carried into the next iteration.

The SM facilitates the team's contribution to ART Sync (Scrum of Scrums plus PO Sync) and drives cross-team dependencies to resolution or escalation.

Predictability equals actual business value delivered divided by planned business value, and is calculated per PI rather than per iteration.

Built-in quality is sustained through XP practices (CI, TDD, refactoring) and a Definition of Done the team keeps current.

The PI closes with the Inspect & Adapt event, comprising the PI System Demo, quantitative measurement, and a retrospective/problem-solving workshop.

The PI System Demo is the ART-level integrated demonstration where stakeholders assess whether the solution met the PI objectives.

Quantitative measurement compares planned versus actual PI objectives and business value, producing the ART predictability metric.

The problem-solving workshop applies root-cause techniques (5 Whys, fishbone) and generates improvement backlog items with owners.

Improvement actions from I&A are entered into the ART backlog and planned into the next PI.

The SM ensures the team prepares for the System Demo, contributes improvement items, and updates status and business value of its PI objectives.

Velocity and predictability trends from the completed PI set the capacity baseline for the following PI Planning.

Risks left open through the PI are reviewed at I&A and either closed or carried forward with updated ROAM status.

AI in the SM role is augmentation, not replacement: AI drafts, summarizes, and surfaces patterns; the SM decides and remains accountable.

Common applications include meeting summaries and action items, agenda drafting, retrospective theme clustering, backlog quality checks, and dependency pattern detection.

Prompt quality drives output quality: role, context, constraints, desired format, and examples must be specified.

AI output must be human-validated before being shared with the team or stakeholders, because hallucinated content is a known failure mode.

Confidentiality rule: no customer data, credentials, or sensitive IP enters a public AI tool unless the organization has approved the tool and its data handling.

AI-driven analytics (sentiment, cycle time, cumulative flow) can flag team-health signals, but such metrics must never be used to evaluate individuals.

Bias review is required: models can encode historical bias, so the SM checks AI outputs for equitable treatment of team members.

Responsible AI - transparency, accountability, privacy, and fairness - is the governance lens applied to all AI use in the ART.