Published 20/01/2026 By nAG
In a previous post, we described the messy middle — the space between legacy systems and full modernization where many enterprises operate. It’s a place defined by risk, constraint, and trade-offs.
Understanding the messy middle is the first step. A harder challenge is knowing how to make good decisions once you’re there.
Enterprises don’t fail in the messy middle because they lack ambition. They struggle because the decisions are subtle, interconnected, have potential negative ramifications, and are sometimes irreversible.
One of the most common mistakes enterprises make is assuming that progress requires immediate change.
In reality, some legacy numerical components are stable precisely because they’ve been left alone. Changing the wrong part of a system — even with good intentions — can introduce more risk than it removes.
The key decision is not what can be modernised, but what should remain untouched for now. Continuity, in this context, is an active choice, not a failure to act.
Compliance risk is often misunderstood as a property of age: old code equals high risk. In practice, the risk usually lies elsewhere.
Unsupported numerical software becomes a compliance problem when it is opaque — when assumptions are undocumented, behaviour is poorly understood, and accountability is unclear. Simply replacing code doesn’t automatically resolve these issues.
A better question is not “Is this software old?” but “Can we explain, justify, and assure its behaviour?” That shift in thinking often changes where effort should be applied.
Modernization is widely accepted as the long-term goal. The danger lies in treating transition as a single event rather than a sequence of controlled steps.
Enterprises get into difficulty when transition is rushed or framed as a finish line. Successful transitions preserve operational trust — with regulators, users, and internal teams — at every stage.
The decision here is about sequencing: what can move, what must wait, and how to avoid breaking systems that must not fail.
The hardest part of navigating the messy middle isn’t tooling or technology — it’s judgement.
Numerical software behaves differently from general enterprise IT. Small changes can have outsized consequences, where correctness, reproducibility, and assurance matter as much as performance or scalability.
This is where experience makes a difference. At nAG, we’ve spent over five decades building, supporting, and maintaining numerical software used in some of the world’s most demanding environments. Our work spans high-performance computing through to modern AI, giving us a deep understanding of both legacy constraints and future directions.
That perspective helps enterprises make better decisions — not just faster ones.
The messy middle isn’t a temporary problem to be solved once and forgotten. For many organisations, it’s the operating reality for years to come.
The difference between stability and disruption often comes down to the quality of decisions made there — what to change, what to protect, and how to move forward without breaking what already works.
nAG has the experience, capability and know-how to bridge the messy middle transition. If you’d like to know more do reach out — we’d be delighted to talk with you.