Scrum and AI: Partners or Rivals?

Why It Matters

Understanding how AI reshapes Scrum’s bottlenecks helps firms avoid wasteful rapid delivery and focus on outcomes that drive real business value.

Key Takeaways

• AI speeds code creation, but discovery remains human‑driven bottleneck.
• Scrum must stay learning‑focused, not just velocity‑driven in sprints.
• Integrating AI requires stricter upfront specifications and governance.
• Outcome metrics outweigh feature count when evaluating AI‑enhanced sprints.
• Strategic lag grows if validation delays follow rapid AI delivery.

Summary

The webinar, hosted by Eric Neberg and Scrum trainer Alex Balerin, explored whether Scrum and artificial intelligence are allies or adversaries in modern product development. After a quick poll revealed most teams are still experimenting with AI at an individual level, the presenters framed the discussion around Scrum’s core purpose and the profound workflow changes AI introduces.

Key insights highlighted that AI dramatically reduces the cost and time of code generation, shifting the real bottleneck to discovery, requirement definition, and validation—activities that still demand human judgment. When Scrum is applied mechanically, focusing solely on velocity, teams risk flooding the pipeline with features that add no business value. Conversely, a learning‑oriented Scrum, combined with disciplined AI usage, can accelerate delivery while preserving empirical feedback loops.

Examples included a GitHub‑based study showing a 17‑fold increase in code output but only a 30 % rise in releases, underscoring the danger of “being too fast.” The presenters cited Joshua Sen’s logical model to stress tracking resources, engineering KPIs, and, crucially, outcome‑based metrics rather than raw feature counts. They warned that strategic lag—delayed validation of shipped code—can render rapid AI‑driven development counterproductive.

The implication for businesses is clear: adapt Scrum ceremonies to incorporate AI governance, enforce detailed upfront specifications, and prioritize outcome measurement. By rebalancing sprint goals toward validated impact, organizations can harness AI’s speed without sacrificing the empirical learning that makes Scrum effective.

Original Description

AI lets teams ship more software, faster, with fewer people. Some developers conclude that in this context, Scrum is redundant. And they have a point - if Scrum is reduced to meetings for controlling work and two-week Sprints, AI makes it dispensable.

But Scrum is not that.

When building a feature costs almost nothing, the bottleneck is no longer production. It becomes the decision: what to build, for whom, and why now. That's where Scrum - properly applied - becomes more critical, not less.

In this session, PST Alex Ballarin explores how a Scrum Team's work changes when AI enters the development flow:

-What copilot-mode development and agentic AI look like in practice, with concrete examples of what changes for the team.

-How Scrum events and artifacts evolve when delivery velocity multiplies.

-Why demand management and human judgment become the new bottleneck.

-Open Q&A on Scrum and AI.

Whether you've been using AI in your team for a while or are just starting to explore it, this session will give you a framework to integrate it without losing focus on what matters.