AI is becoming the engine room of digital business. As that journey continues, it is ramping up the volume, complexity, and intensity of the workloads cloud infrastructure needs to support. According to Gartner, there will be a five-fold increase in AI-related cloud workloads by 2029 – which will consume half of all compute resources. In just a few years’ time, only those with a truly modern cloud foundation will be able to keep up in the race for AI-driven supremacy.
This is ramping up the urgency for enterprise cloud migration. Over the past decade, the cloud was a good option to choose. Today, AI has made it a necessity. In this new era, businesses need the right architecture, security model, and agility to capitalise on the AI opportunity. Every day spent operating on legacy architecture widens the gap, as competitors deploy AI faster to gain operational and commercial advantage.
Many organisations just aren’t ready for AI
Over the past decade, many organisations have modernised their applications and embraced cloud-native architecture – and they’re now well positioned for the surge in AI-driven workloads. But a large proportion are not in this position. Across industries, the same two patterns keep emerging:
Some organisations stayed on-premises, held back by monolithic systems, fragile applications or years of accumulated technical debt.
A second group of organisations did move to the cloud, but without modernising first. This lift-and-shift approach simply transferred legacy issues into a new environment. In many cases, these organisations are spending more on their infrastructure than they were before, without unlocking any real benefits.
These two groups now face the same challenge: their environments cannot support AI at scale. AI needs elasticity, modern security patterns, clean integrations, and real-time access to data. Most legacy or superficially migrated applications simply can’t provide that. To drive scalable innovation with AI, organisations need to switch to a modernisation-led re-architecture of their cloud applications.
The starting point: understanding what you have
The first step on this journey is to create a clear inventory of the entire application estate. It is a mistake to overlook or downplay the importance of this step – as the absence of an application inventory is a common reason behind modernisation and migration failure.
Businesses should be looking to group their applications into three core categories. First, their business-critical applications that directly affect revenue or customer experience. The second group is formed of business-enabling applications that support operations and processes. Finally, there are corporate and support applications used across HR, finance, and internal teams.
Splitting the estate into these groups will inform the decision-making process – for instance, highlighting high-priority applications, or identifying some that are no longer needed. At this stage, it’s important to fully understand the foundational business logic within applications and ensure it is protected while addressing the architectural issues.
Automation can accelerate this assessment phase, reducing the time it takes to analyse architectures and identify problems. It’s also worth asking the organisation’s transformation partners to identify opportunities to remove underutilised licences or infrastructure, opening up the path to a self-funded modernisation.
Choosing the right modernisation strategy
Once organisations know what they have, they need to identify the right modernisation strategy for each application, making a choice between the five “R’s”:
Rehost – a simple lift to cloud infrastructure
Re-platform – minor upgrades, or containerisation
Refactor – optimising parts of the code for cloud-native services
Rearchitect – fundamentally redesigning the application
Replace – with a SaaS or off-the-shelf solution
Across these options, each application should be matched to a strategy based on its criticality, cost profile, performance needs, technical debt, and long-term value to the business. The goal is not speed, but sustainability: modernisation should leave applications more resilient, flexible and secure than before.
In some cases, it makes sense to modernise an application on-premises first, then migrate it to the cloud once it is stable. This reduces risk, while still moving the organisation towards a cloud-native foundation.
Oversight and structure are key
Once modernisation is underway, migration needs to be anchored by strong governance. A central structure – such as a Cloud Transformation Office or a Cloud Centre of Excellence – can provide the oversight needed to keep scope under control, manage risk, and maintain business alignment. Organisations that have never attempted a transformation of this scale should draw on the advice of expert partners to identify an appropriate governance model and steer the programme towards success.
With strong governance in place, the migration itself should follow a structured path. The most effective programmes take a batch-based approach, grouping similar workloads so teams can accelerate progress and reuse proven patterns instead of treating each application as a one-off. Automation should underpin this entire process, providing consistency, accuracy and speed across every stage of the transformation.
Modernise or fall behind
Some organisations claim to be “in the cloud”, yet are seeing few of the benefits. Others haven’t moved at all. But as AI drives an explosion in workload volume and complexity, the cost of inaction is rising rapidly. The organisations that will thrive in the AI era are those that modernise effectively, migrate with discipline, and treat cloud not as a data centre – but as a platform for innovation. Modernisation helps you build sustainable and optimised long-term cloud-native solutions, creating the foundation AI needs to deliver lasting value.
Modernisation-led migration is no longer a technical choice. It is the foundation for unlocking the value of AI – and the dividing line between the businesses that will compete effectively in the next decade, and those that won’t.
