Modernization Keeps Getting Stuck. Here are 4 Paths Out.

In my experience, most enterprises have already modernised and moved the “easy” workloads. The newer applications, the less-embedded systems, the stuff where the risk was manageable, which usually equates to around 20-30% of their estate.

What remains is the critical, complex and undocumented core. The systems that have been running for 20, 30, 40 years, where the people who built them have retired, and where knowledge of them exists in just a handful of heads, if any.

This is where many organisations find themselves stuck. It’s not that they don’t want to modernise or understand the importance of it, it’s that they don’t fully understand what they have, which makes the risk of something going wrong during a modernisation initiative unacceptably high.

The Data Backs this Up

Ensono’s recent State of IT Modernisation 2026: Enterprise Reality Check report, surveyed 500 IT leaders across the UK and US. The UK findings confirm what I’ve been hearing in conversations with technology leaders for years:

67% have had to pause, scale back, or abandon modernisation initiatives, and 77% exceeded budget (both these figures are higher than their US counterparts).

These numbers make sense because modernisation is never simple for UK enterprises. The regulatory environment is heavier, the legacy estates are often older and, let’s face it, our talent pool for “legacy” skills is much smaller and shrinking fast.

AI is adding urgency, but it shouldn’t force bad decisions. Our research found that 54% say AI is accelerating modernisation, while 78% say their legacy systems are more important than they were two years ago. That’s not a contradiction. The valuable data and business logic AI needs often sits inside those older systems. The challenge is working out what to retain, what to modernise, and what to retire.

Transformation Failures Have Made Us All More Cautious

The UK market has seen too many big-bang programmes go wrong. Migrations that ran over by years. Lift-and-shift projects that uncovered complexity nobody had anticipated. Consultants who promised transformation and delivered chaos.

And in regulated industries, every change must pass compliance scrutiny — FCA, DORA, ACORD, operational resilience rules — adding time, cost, and risk to every modernisation decision.

So many technology leaders delay making the hard moves, choosing instead to wait for a burning platform before they act.

I understand the caution, but delaying modernisation is a decision that comes with its own costs. Our research found that 97% of IT leaders say talent shortages have already raised their modernisation costs, and that the pressure is only increasing.

The commercial impact of standing still can’t be underestimated. The ever-shrinking pool of legacy skills, reduction operational resilience, increase in technical debt, impact on future AI adoption and the increasing cost of change are all topics that come up regularly in board conversations

The longer you wait, the scarcer the skills become, the more technical debt accumulates, and the harder it is to adopt AI at scale.

There’s No Single Right Answer. That’s the Point.

The industry loves a simple narrative like “Get to cloud,” “Protect the mainframe, ” and of course, “AI changes everything”, but the reality is messier because every organisation is different.

Some have no plans to leave their legacy platforms, and they’re right not to.

We recently worked with an insurance client running IBM Power systems with 6 million lines of RPG code processing 30 million requests daily. They didn’t want off. They wanted to understand what they had and prove AI could work in their environment.

Others want off entirely, and they’re right, too.

A major UK retailer we work with has made it clear: they want to be off the mainframe within five years. It’s part of their technology strategy because it’s best for their business. Our goal is to help them get there safely, without breaking anything critical along the way.

In the end, it’s not about which approach is “correct”, it’s about making the decision based on your business, risk profile, and actual estate, and not based on someone else’s ideology.

The Four Paths Out

When we meet with IT leaders who are trying to work out their modernisation strategy, we keep coming back to four things:

• Start with visibility

This is why so many organisations are stuck. The documentation doesn’t exist, the knowledge is in people’s heads, and nobody fully understands the dependencies. You can’t plan around what you don’t understand. Without visibility, any plan is really just a guess.

• Focus on the business outcome

Boards don’t care about lines of code. They care about speed to market, operational efficiency, and what the estate means for company valuation. Lead with that.

• Ignore the ideology

“Cloud-first” and “protect the mainframe” are both ideological positions. The right answer depends on your applications, your business, your risk profile. Be wary of anyone pushing a single approach for every situation.

• Find a partner who isn’t conflicted

If your vendor only succeeds when you stay on their platform, they’ll never tell you when it’s time to leave.