Scrum Guide Expansion Pack 2025 – my first impressions

Yesterday, I participated in the official online release of the Scrum Guide Expansion Pack developed by Jeff Sutherland, John Anthony Coleman and Ralph Jocham. Since then I’ve read the new guide and here are my first impressions - the good stuff and the gaps I still see.

The idea behind the expansion pack is simple: keep the core framework lean and recognisable, but layer on extra explanations and modern perspectives so teams get more of the how and why without having to learn a brand-new method.

Where the 2020 guide was held to a minimum, the new pack stretches across roles, artefacts, events, theory, scaling, AI and an appendix on adaptive strategy. It’s declared community-driven, meaning updates will be released much like software patches - small, frequent, and based on real-world feedback.

Two forces have pushed the authors to add more substance:

• Rising complexity. Products now live in ecosystems shaped by IoT, climate regulation and hyperspeed competition. Simple practices meet heavier context.
• Exploding AI tooling. Machine-learning assistants can boost (or derail) empiricism.

The pack therefore adds dedicated content on stakeholder involvement, outcome measurement, AI as a team member and “multi-Scrum-team products.”

In short, the pack tries to bridge the gap between a tidy 2020 world and today’s messy reality.

Things I love about the Expansion Pack

Stakeholder & Supporter are first-class roles Stakeholders are finally treated as core players, and Supporters are the people who actively remove organisational roadblocks. Clearer accountability means fewer “management-isn’t-helping” complaints.

AI can be a team-member - provided a human stays in the loop The pack admits that AI agents can accelerate discovery, backlog grooming and pattern spotting while insisting that humans remain the final decision-makers. That opens the door for GPT-style assistants without surrendering responsibility.

Two new commitments: Definition of Output Done & Definition of Outcome Done “Done” is split into technical quality and proven value. Focus shifts from “features shipped” to “problems solved,” giving developers and business folks a shared language.

A solid theoretical foundation: Approximately half the document revisits Complexity, Lean, Systems and Product Thinking, Discovery, People & Change, and more. Those pages are gold when you need to justify empiricism, small experiments and sustainable pace to senior leadership.

Scaling advice that’s both strict and pragmatic One Product Goal, one Product Backlog and one baseline Definition of Done across multiple teams—and a frank warning not to scale until a single team runs smoothly. Simple guard-rails that help you avoid SAFe-scale over-engineering.

Optional artefacts: Product Vision, Acceptance Criteria & Outcome Criteria: The expansion pack proposes lightweight, optional artefacts that sit above or beside the Product Backlog. A clearly articulated Product Vision anchors strategy; Acceptance Criteria give teams a shared understanding of “good enough” for each backlog item; and Outcome Criteria focus everyone on the real-world impact.

These additions tighten the feedback loop between long-range intent and day-to-day delivery—without forcing extra ceremonies on every team.

Things I still miss

A strategic leadership role Many companies have merged or even scrapped the Scrum-Master title. The pack keeps the Scrum Master as change agent but never defines a role with the organisational mandate to connect product vision, culture and structure across teams. When the Scrum Master is downsized, teams still need a strategic leader to protect empiricism and ensure top-level alignment.

A fully fleshed-out cadence bridge from Product Vision to Sprint Goals The new optional Product Vision is great, and Sprint Goals still anchor each iteration—but the pack offers no roadmap for how a lofty vision is systematically broken down into Product Goals and then cascaded into coherent Sprint Goals week after week. Without explicit techniques teams risk writing Sprint Goals that feel random, disconnected or purely technical, leaving stakeholders unsure whether the sprint work is actually moving the product toward its stated north star.

A maturity model and tailored support Self-management is pitched as the ideal, but there’s no shared map of where a team is today or how needs change over time. Novice teams often require hands-on facilitation and firm guard-rails; seasoned teams need greater autonomy and subtler coaching. Lacking a maturity lens, we risk over-helping some teams and smothering others with process.

Language for quantum leaps Continuous-improvement focus is great, but sometimes it isn’t enough and a quantum leap is needed - a deliberate, larger-scale shift in architecture, market or mindset. The pack offers no vocabulary or guidance for that scenario, making it easy to keep polishing when the market demands a new chessboard.

Psychological safety as an explicit practice The Scrum Values, Courage, Openness, Respect, Commitment and Focus, only thrive when people feel safe enough to speak up, surface risks and admit mistakes. The pack nods to “People & Change,” but offers no concrete rituals for building psychological safety. Without deliberate attention to safety, transparency and inspection risk becoming performative rather than real.

Conclusion

The Scrum Guide Expansion Pack 2025 is the long-awaited leap that sharpens roles, puts outcomes first, and welcomes AI, while reinforcing the 2020 guide with a solid theoretical foundation. It’s a valuable add-on for teams already fluent in core Scrum, but only one stop on the road to true organizational agility. Download it, test it, and tell the community what lands and what misses.

These are just my initial impressions after a first read-through; I’m eager to dive further into the details and try the new ideas with real teams.

Finally, my thanks to the authors for the massive effort behind the Expansion Pack and for embracing a community-driven approach that lets Scrum evolve in genuine agile increments instead of three-year drops.