Thoughtworks retreat explores AI’s Agile software future

Thoughtworks held a participant-led retreat on the future of software development, marking the 25th anniversary of the Agile Manifesto and putting artificial intelligence at the centre of discussions with clients and industry practitioners.

The event, called the Future of Software Development Retreat, brought together senior technology leaders for an unconference-style programme. Thoughtworks said the agenda focused on how software development practices may change as organisations adopt AI across the product and service lifecycle.

The retreat used an open format based on Open Space Technology principles. Participants shaped the sessions themselves and joined discussion groups rather than attending scheduled presentations.

Agile anniversary

The Agile Manifesto, first published in 2001, set out values and principles that influenced modern software delivery. Thoughtworks linked the retreat to the anniversary as a point of reflection on how those ideas apply in current conditions.

Martin Fowler, Chief Scientist at Thoughtworks and a signatory of the original manifesto, said the shift created by AI warranted the same level of scrutiny as earlier changes in software development.

"The Manifesto for Agile Software Development represented a statement from several folks who were exploring a striking change in thinking about how software should be built," said Fowler. "With this event, we are gathering those who are exploring the change AI brings to the very foundations of how we build software."

Responsible use

Thoughtworks framed the discussions around responsible AI adoption in engineering organisations. It emphasised maintaining collaboration and engineering judgment as teams increase their use of AI tools.

Participants discussed the role of technical standards in AI-assisted development. Topics also included how organisations can manage quality when AI systems generate or modify code and documentation. Thoughtworks also highlighted the need for governance structures that align with the pace of AI adoption within delivery teams.

The company said conversations covered how teams can maintain what it called technical excellence. It also raised questions about how roles may change, including how product, engineering, and operations groups will divide responsibilities as AI systems take on routine tasks.

Platform demonstration

During the retreat, Thoughtworks demonstrated AI/works. The company described it as an agentic delivery platform.

Thoughtworks said the platform synchronises AI agents across discovery, delivery and operations. It positioned the demonstration as a way to ground discussions in practical examples of how AI systems might operate within software development workflows.

The company said the demonstration sparked conversations about governance, trust, and scale in AI-enabled software development. These themes have become more prominent as organisations move from individual productivity tools to AI systems that operate across multiple stages of delivery.

Rachel Laycock, Chief Technology Officer at Thoughtworks, said the retreat aimed to create space for debate rather than present fixed guidance. "This retreat is not about presenting finished answers. It is about creating space for honest exploration at a pivotal moment for our industry. As AI becomes embedded in software delivery, we need to rethink practices and roles while staying grounded in the human-centred values that made Agile transformative."

Industry questions

Across the software sector, AI tooling has expanded from code completion to systems that propose designs, generate tests, review changes and assist with incident response. Organisations have begun applying these tools in regulated environments and large engineering estates. This has raised questions about auditability, accountability and the reliability of outputs.

Thoughtworks highlighted governance and trust as key areas for discussion. These concerns often include how teams validate the output of AI systems, how they document decisions, and how they manage risks when models change or operate with incomplete context.

The retreat also addressed how Agile principles translate to an AI-first world. Thoughtworks described this as an examination of what needs to change in how teams organise work and measure outcomes, while maintaining the core ideas of Agile delivery.

Thoughtworks said it expects insights from the retreat to inform future articles, podcasts and videos. The company said it plans to extend the discussion beyond the event through its published content.