Scrum.org CEO Dave West on Scaling Scrum, AI and Team Autonomy

Scrum’s value depends less on enforcing its rituals than on whether teams use them to learn and deliver useful work. That’s the view of Dave West, CEO and product owner at Scrum.org, who sees a familiar failure in organizations that say they use Scrum while following plans imposed from above. His advice to Scrum Masters is direct: help the team succeed; don’t turn the role into rule enforcement.

In an interview with Techopedia, West discussed the difficulty of aligning Scrum across large organizations, the role artificial intelligence might play in team workflows, and the risks of drawing confident conclusions from incomplete data. His perspective is shaped by work in software development, research, consulting and product management, including leadership roles at IBM/Rational, Forrester Research and Tasktop.

Scrum works when teams can adapt

West describes Scrum as a way to deliver value in increments, particularly when the work is complex and teams need feedback to decide what to do next. Its events, artifacts and accountabilities provide a structure for that work. They aren’t meant to dictate every team decision.

The Scrum Guide is 13 pages, West said. Its core ideas are to set a goal, meet to maintain focus, review work to get feedback, and come together to improve. How a team puts those ideas into practice is largely up to the team.

That distinction helps explain why some teams experience Scrum as rigid. The structure can be taught with clear definitions and routines, West said, but applying it well requires judgment. Teams need to understand the intent behind the practices and then adapt their approach to the work.

When an organization holds Scrum events but leaves teams no say in the work, the framework can become a performance. West has written about this pattern under the name “Water Scrum Fall”: a team goes through Scrum’s motions while being handed work and decisions made elsewhere. In that setting, the meetings don’t create the empirical feedback or autonomy that Scrum is meant to support.

West’s criticism is not that structure has no place. A team may need a more consistent approach while it learns how to work together; later, it may need less direction. The Scrum Master should help the team find that balance, with the goal of delivering value rather than demonstrating compliance.

Scaling means aligning the whole organization

Growing Scrum beyond a single team creates organizational choices that can’t be solved by copying team practices at a larger scale. West points to decisions about product focus, how teams are arranged, and what behaviors organizations reward. If those choices pull in different directions, teams may struggle to produce outcomes even when their own routines work well.

His systems view puts the emphasis on the organization as a whole. Improving one team in isolation can leave it constrained by dependencies, incentives or priorities elsewhere. Leaders need to consider how the parts fit together and optimize for the larger system.

That can be harder than announcing a new direction. Leaders may want to reduce hierarchy, focus on outcomes and let teams make intelligent choices, West said. But changing how people approach work and relate to one another takes time. West sees examples of agile practices in parts of organizations, but said he doesn’t see them consistently across entire companies, including large technology businesses.

He expects the nature of knowledge work to keep changing as AI tools take on repetitive tasks. If people have more time for creative work, collaboration becomes even more central: teams will need to work through problems together, much as creative groups do in fields such as filmmaking. West sees that shift as a reason to think about organizational design as well as the tools themselves.

AI as a working aid for teams

West expects AI, particularly large language models, to act as a supporter or coach within team workflows. Rather than defining a new framework, these tools could offer suggestions: help a team work on its Sprint Goal, find a canvas for exploring a problem, or point out that backlog items are written as tasks instead of outcomes.

Those examples describe a possible assistant role, not a claim that AI can set a team’s priorities or judge whether its work matters. West hopes such tools will help people get past tedious work and spend more time on customer needs, product goals and outcomes. He’s also clear that the future he describes is still uncertain.

He doesn’t see a new agile framework on the horizon. Instead, he expects existing approaches to evolve through extensions and combinations with design thinking, lean startup methods, coaching, modern management practices and clearer alignment between objectives and key results and the outcomes teams pursue.

For organizations evaluating AI in development work, West’s comments point to a practical question: does a tool help a team understand a problem and make better decisions, or does it add another process layer? The examples he gives focus on assistance with specific tasks and feedback. They leave open how well these systems will handle context, and how teams should check their suggestions.

Distributed work and data-informed decisions

West is also interested in how teams collaborate across locations. The shift to distributed work in 2020 challenged the assumption, common in parts of the agile community, that teams had to share one workplace to work effectively. West says experience has shown that distributed teams can deliver for customers.

That doesn’t make collaboration automatic. Teams still need ways to share information, discuss work and learn from one another. West’s point is that location alone doesn’t determine whether a team can work together; the practices and tools supporting the work matter too.

He sees data as another area of change. Collecting information about software work used to take substantial effort. Today, organizations can gather more data and draw on analytics tools to inform decisions. That increased access can help, but the quality and context of the information still matter.

West illustrated the change with a survey he had recently run about how organizations think about products: whether they have a product model, can identify their products, and distinguish product work from project work. It received around 4,500 responses in a week. For him, that shows how widely accessible research has become. Surveys and digital tools let more people gather information without relying solely on large research firms.

Accessibility, however, doesn’t guarantee reliable conclusions. A survey may have a biased sample or omit important context. The person interpreting its results can also bring assumptions to the analysis. West warned that it’s easy to find data that appears to support an argument without establishing that the argument is correct.

He wants teams to apply critical thinking alongside data analysis, including attention to bias, ethics and context. Data can shape which groups benefit from a product and which are overlooked. A feature that appears popular may still have consequences its builders haven’t considered. West’s concern is that teams can mistake a strong signal or an appealing capability for a complete understanding of its effects.

AI adds to that challenge. As people turn to AI tools for answers, West argues that they need to preserve their own critical judgment. Data can improve decisions, but it can also reinforce assumptions when teams don’t ask where the information came from, what it leaves out, or how it will be used.

From services work to a product business

West’s experience at Tasktop offers a concrete example of the tradeoffs involved in product development. The company continued providing services to bring in revenue while building toward a product business. A clear vision and roadmap helped connect that client work to the longer-term direction, and product management was introduced relatively early.

Tasktop then received $12 million from Yaletown and Austin Ventures. West said the funding went toward expanding product management and engineering, building on work already underway to deliver products consistently.

The transition still involved difficult choices. A client could offer a large payment for a custom feature, creating immediate pressure to prioritize that request over the product roadmap. For a company paying its bills through services, the temptation is real. West said the team needed to stay focused on its vision while managing that tension.

He credited founder Mik Kersten with helping maintain that direction. Kersten wrote Project to Product and created Eclipse Mylyn. West also pointed to the wider team’s role in supporting the shift. The account illustrates a broader product-management challenge: revenue from individual client requests can sustain a business, but a steady stream of custom work can make it harder to build a product around a consistent direction.

Leadership built through learning and kindness

For technology leaders trying to make change, West’s advice centers on three habits: build a network, make time to learn, and ask whom you’ve helped rather than simply what you’ve completed. He recommends putting reading and learning directly into the schedule.

He offered the example of a Scrum Master who uses a script to collect material from publications such as The New York Times and sets aside an hour each day to learn. The details are less important than the routine: professional development is easier to sustain when it has a regular place in the working week.

Helping other people can also strengthen a professional network while exposing leaders to new ideas. West said conversations with people pursuing roles as entrepreneurs, product owners or CEOs can teach him something, even when he’s tired or busy.

He connects that approach to kindness. Instead of dismissing a colleague’s idea, he suggests asking why they see it that way or acknowledging that the perspective is different. For West, being approachable and asking questions helps people work through disagreement without shutting down discussion.

That advice ties back to his view of Scrum. A framework can give teams a shared way to focus, inspect their work and improve, but it can’t substitute for the human conditions that make collaboration possible. Teams need room to adapt, leaders need to align the wider organization, and data and AI tools need thoughtful use. For Scrum Masters, West’s measure remains the team’s success in delivering value.

About Dave West

Dave West is CEO and product owner at Scrum.org. He co-authored The Nexus Framework For Scaling Scrum and Head First Object-Oriented Analysis and Design, and led development of the Rational Unified Process at IBM/Rational. He later managed Ivar Jacobson Consulting in North America and led Forrester Research’s software development and delivery practice. Before joining Scrum.org, he was chief product officer at Tasktop.

Background on Scrum adoption: Scrum in project management. Related definitions: artificial intelligence and large language models.