Being agile in a world of AI

AI is here but it's a long way off disrupting human team dynamics, says Michael O'Reilly

Pro

In association with NTT Data

When the founders of the Agile Manifesto met back in 2001 in a remote Utah ski resort, little did they know the impact four simple values and 12 principles would have on the world of software development.

The agile methodology now drives the world of software development, with published statistics indicating adoption is over 80% globally and growing.

The most interesting fact about the manifesto is while its intention was “uncovering better ways of developing software,” it did not mandate tools or technology to help you achieve this. The manifesto principles focus on enabling people to deliver better software by supporting them to get the job done. By association, they will self-organise to maximise their ability to deliver faster with greater alignment to the business’ actual needs. Unsurprisingly, they would turn to technology to achieve these outcomes by building better collaboration tools, deployment tools, test tools and new practices through DevOps, etc. While the evolution of Agile is driven by technology, they probably could not have predicted the potential impact of AI on the industry. The question is what impact AI will have on agile software development.

AI will transform our world with the latest developments in generative AI, bringing the debate to the software development community. Developments in natural language processing and large language models can be turned to any discipline that involves text and language. Agile software development is a target for this technology. Business requirements are defined, elaborated, detailed, implemented, tested and deployed through various text forms. Building an AI model to tackle the above is possible with sufficient sample data and it is safe to assume there is plenty of data if 80% of the world’s projects are agile.

Impact

We can already see AI’s impact on software development in various areas of the software development life cycle (testing tools; code quality; vulnerability and security fixes; code generation including low/no code, self-healing code etc.). The industry is adopting AI quickly, delivering solutions that help people build better solutions. As AI technology evolves, will this role of helping people move to replace people in the life-cycle?

Already, we are seeing Generative AI develop and manage business requirements, build content for systems, write code, develop test cases, and more. Will this evolution impact the roles typically found in an agile team?

The simple answer is yes. No doubt AI will and is impacting all the roles in an agile team. For example, in the scrum master role, AI can better analyse and compare data, define agendas, schedule meetings, track metrics, and complete routine tasks. However, the scrum master role is much deeper than the mechanics, where more human skills are required to mentor and motivate the team and negotiate priorities and dependencies internally and across teams. Currently, these things are beyond the capabilities of AI but who knows what the future will bring.

Reflecting on the Agile Manifesto, its principles are adaptable to the use of technology and continuous learning to improve the software delivery process. As such, AI is just another type of technology that helps an agile team achieve better outcomes and is in line with the manifesto’s principles.

To that end, technology that helps people and teams is beneficial and should be embraced. AI will help every team member deliver working software faster, which is the central theme of the Agile Manifesto. Maintaining the flow of value to the business and fostering a balance of roles across an agile team is essential to team dynamics, with no one part dominating, AI or otherwise.

The debate on responsible AI will continue with determining the impact on agile software development. The Agile Manifesto may be adaptable in its current form to take on AI just like it has adopted many other technologies. Undoubtedly, it will help accelerate our ability to deliver better software in collaboration with our AI counterparts.

Michael O’Reilly is CTO, international business, with NTT Data

Read More: NTT Data