AI factories promise to turn intent into working software and products with minimal human hands. We’ve built them in greenfield and brownfield settings, and the reality is more interesting than the pitch.
This session compares the factory patterns we’ve run in practice. They range from staged pipelines that move from issue to spec, build, review and release, to setups driven by a brief, a mission or a digital twin, where the factory works from product context instead of tickets. For each pattern we look at where it delivers, where it quietly fails, and what it takes to verify results when no human has read the code. A green test suite is not that proof.
We draw on real examples from our own and client work. You’ll leave with a clear view of which factory approaches fit which situation, what they demand from your product organisation, and where the hype outruns reality.
Managing Director at Mayflower.
Björn is the Managing Director of Mayflower and an expert in AI, modern data strategies, and hyper-automation. The approximately 80-member Mayflower team develops tailored software solutions – agilely in teams and in close collaboration with clients. His focus areas include Agentic AI, fast data access, and modern software architecture for sustainable digital innovation.
