Scrum with AI is a lightweight framework that helps individuals, teams, and organizations generate value through adaptive solutions in environments enhanced by Artificial Intelligence.
This guide defines Scrum with AI. It consists of Scrum’s core, extended with AI-specific roles, responsibilities, and principles to ensure ethical, data-driven, and sustainable delivery.
Scrum with AI is an iterative and incremental framework that enables teams to deliver intelligent, adaptive, and responsible solutions by combining human creativity with AI capabilities.
The core framework of Scrum remains unchanged. Scrum with AI does not alter the foundational roles, events, artefacts, or principles of Scrum. Instead, it extends Scrum by embedding AI capabilities and human judgment together to enhance decision-making, learning, and value delivery.
This approach ensures that while Scrum continues to provide structure and discipline, AI acts as an amplifier of insights, speed, and adaptability not a replacement for human accountability.
Scrum with AI is founded on empiricism and lean thinking, and is further strengthened by augmentation thinking.
Knowledge comes from experience, augmented by insights derived from AI models and data systems. Decisions are made based on observed outcomes, with AI enhancing the speed and depth of learning without replacing human judgment.
Transparency in Scrum with AI goes beyond visibility of work items. It includes clarity on data sources, model behaviour, assumptions, and limitations. Teams must ensure that stakeholders understand how AI-driven decisions are made and what factors influence them. The intent is to build trust and enable informed decision-making rather than blind reliance on AI.
Inspection involves continuously evaluating both the product and the AI systems powering it. This includes reviewing model accuracy, detecting bias, and assessing real-world performance. Inspection is not limited to outputs but extends to data pipelines, training processes, and feedback loops. The goal is to uncover hidden risks and improvement opportunities early.
Adaptation in Scrum with AI requires responding not only to changing requirements but also to evolving data and model behavior. Teams may need to retrain models, update datasets, or adjust algorithms based on new insights. This ensures that the system remains relevant and effective in dynamic environments. The intent is to create systems that learn and evolve continuously.
Focuses on reducing waste not only in processes but also in unnecessary data usage, model complexity, and computational cost. The intent is to create efficient systems that deliver maximum value with minimal overhead.
Augmentation thinking emphasizes that AI is designed to enhance human capability, not replace it. Humans provide context, judgment, ethics, and strategic direction, while AI contributes scale, speed, and pattern recognition. The intent is to create a symbiotic relationship in which human and artificial intelligence amplify each other to achieve superior outcomes.
In addition to the original Scrum values, Scrum with AI introduces the following extended values:
Teams take ownership not only of delivery but also of the impact created by AI systems. This includes accountability for outcomes influenced by automated decisions and ensuring alignment with organizational intent.
Every role within the Scrum with AI framework is accountable for its contributions, including AI-driven outputs. This reinforces clarity in ownership, especially when decisions are supported or influenced by AI systems.
Teams remain continuously aware of ethical implications in data usage and AI behavior. This includes proactively identifying bias, ensuring fairness, and safeguarding user trust. Ethical awareness is not a checkpoint—it is a mindset embedded in everyday work.
The Scrum with AI Team comprises professionals focused on delivering a “Done” Increment each Sprint.
Responsible for maximizing the value of the product, including AI-driven outcomes.
In the age of AI, the Product Owner evolves from a backlog manager to a value maximizer who leverages AI to make better, faster, and more informed decisions.
Responsibilities:
Responsible for establishing Scrum as defined in this guide and ensuring flow efficiency.
In Scrum with AI, the Scrum Master augments their capabilities using AI to improve team effectiveness, decision-making, and continuous improvement.
Responsibilities:
Professionals who build, integrate, and maintain both software and AI components.
In the age of AI, the term “Developers” extends beyond coding. These are solution builders who collaborate with AI systems to design, create, and evolve intelligent products.
Responsibilities:
Note: While the term “Developers” is retained to preserve alignment with Scrum, organizations may choose to adopt alternative terms such as “Solution Builders” or “Intelligent System Creators” to reflect this evolution.
Ensures that AI systems are effectively designed, selected, trained, and integrated.
Responsibilities:
Ensures responsible use of data and AI.
Responsibilities:
The Strategist bridges organizational vision with AI capabilities. This role is typically played by leaders who guide how AI aligns with business strategy and long-term transformation goals.
Responsibilities:
The intent of the Strategist role is to ensure that Scrum with AI does not operate in isolation, but is deeply connected to enterprise direction and impact.
An ordered list of everything needed to improve the product, including:
Selected items for the Sprint, including AI experiments and validations.
A usable outcome that may include:
Describes a future state including intelligent capabilities.
Defines why the Sprint is valuable, including learning goals for AI.
Includes:
A fixed-length event where value is created, including AI experimentation.
Includes:
Focuses on:
Includes:
Includes:
AI systems are not static; their performance evolves over time. Continuous model evaluation ensures that models remain accurate, relevant, and aligned with real-world conditions. This includes monitoring for model drift, validating outputs against expected results, and recalibrating when necessary. The intent is to treat AI models as evolving assets that require constant attention, similar to product increments.
AI outcomes are only as reliable as the data on which they are built. Data quality monitoring focuses on ensuring the completeness, accuracy, consistency, and timeliness of data. This practice emphasizes identifying anomalies, missing data, or bias in datasets early. The goal is to build trust in AI systems by maintaining high-quality data pipelines.
AI should augment, not replace, human decision-making. Human-in-the-loop validation ensures that critical decisions, edge cases, and ambiguous outputs are reviewed by humans. This creates a balance between automation and accountability. The intent is to combine computational efficiency with human judgment to reduce risks and improve outcomes.
Responsible AI checkpoints are built into the workflow to ensure that ethical, legal, and societal considerations are continuously addressed. These checkpoints validate fairness, transparency, privacy, and compliance. Instead of being a one-time activity, responsibility becomes an ongoing discipline integrated into the delivery process. The aim is to prevent harm while enabling innovation.
Scrum with AI is not about tools or techniques.It is about enabling organizations to build adaptive, intelligent, and responsible systems that evolve with time.
Scrum is a framework originally created by Ken Schwaber and Jeff Sutherland. The core Scrum framework, its principles, and structure remain unchanged and continue to belong to its original creators.
Scrum with AI is an extension guide that builds upon the foundation of Scrum by integrating AI capabilities and strategic perspectives for modern organizations.
The authors of Scrum with AI are Venkatesh Rajamani and ArunVignesh from tryScrum.
Leadership Education
Scrum Master
Product Owner
Coaching Workshop
Kanban Training
Business Agility
Design Thinking
Developer
