Blog
Legacy modernisation without the big bang
7 min readMatthew Ryan
Big-bang legacy replacements have a poor track record and an even worse reputation. The good news is that modern engineering practice — combined with AI-assisted discovery — makes incremental modernisation genuinely viable for systems that used to look untouchable.
The pattern we keep returning to: reverse-engineer the behaviour you actually depend on, wrap it behind a stable interface, then replace it slice by slice. Each slice ships value, retires risk, and gives leadership a credible reason to keep funding the program.
AI accelerates the discovery and test-generation work that used to take months. It does not replace the judgement calls about what to keep, what to retire and what to redesign — that's still where senior engineers earn their keep.
More from the blog
Moving beyond AI pilots: what production-grade looks like
Most AI proofs-of-concept stall before they hit production. Here's the operating model we use to get them across the line.
Architecture decisions that survive turnover
Lightweight decision records are the highest-leverage artefact an architect can produce. Here's the format we use.