Stay, Move or Modernise? A Practical Framework for AI-Ready TMT Estates

As leaders across telecommunications, media and technology (TMT) seek to productionise AI, they must first consider which workloads are ready to support it, which need to change and which no longer justify the cost of keeping them.

However, many organisations are managing a mix of VMware environments, cloud platforms, legacy applications, data workloads and operational systems. Some continue to perform well while others have become expensive to run, difficult to scale or too complex to adapt to new AI, data and automation needs.

A broad migration programme may feel like the obvious answer, especially as commercial models and cost pressures change. But without workload-level clarity, leaders risk moving existing complexity into a new environment.

To modernise with confidence, TMT organisations need a practical way to decide what should stay, what should move, what should be modernised and what should be retired.

Determining workload priority

Cloud migration should not be treated as a single decision. Different workloads support different parts of the business, carry different risks and create different levels of value.

This is particularly true in TMT organisations, where workload decisions are rarely just about infrastructure placement. They are tied to service reliability, customer experience, data monetisation, digital product delivery and the cost of running increasingly complex estates.

In telco, this means understanding which workloads are most critical to network automation, service assurance, customer care, OSS/BSS modernisation, digital voice and enterprise connectivity. A legacy platform may still support an essential service, but it may also increase cost-to-serve, slow down new service launches or limit the use of AI in operations.

In media, workload priorities are increasingly shaped by the shift to streaming, connected TV, personalised experiences and data-led advertising. Platforms that support content storage, rights metadata, audience data, ad decisioning and content supply chains need to be assessed not only for cost and resilience, but also for how well they support faster monetisation and more responsive customer experiences.

For technology organisations, the focus is often on product velocity, engineering productivity, customer support and secure data-driven services. Workloads that underpin SaaS platforms, development environments, usage analytics or AI-enabled product features need the scalability and governance to support rapid change without increasing operational risk.

Across all three sectors, VMware environments support critical applications, data workloads and operational systems. These environments may continue to be the right fit for some workloads, but others may need a different foundation as cost pressures, AI requirements and scalability demands increase.

A workload-led approach helps leaders avoid broad assumptions. It gives them a clearer view of what should remain on VMware, what should be optimised, what should move to AWS and what may need to be replatformed, refactored or retired.

To assess this, leaders can consider five key questions:

1. What business outcome does this workload support?

Every workload should be assessed against the outcome it enables. Does it support revenue generation, customer experience, operational resilience, regulatory compliance, productivity or innovation?

This helps leaders avoid modernising systems purely because they are old or overlooking systems that quietly support critical processes.

In TMT, this distinction matters. A billing platform, content workflow, service assurance tool or customer data system may not always be the most visible part of the estate, but it can have a direct impact on customer trust, revenue and operational performance.

2. Is the cost still justified?

Some workloads become expensive because they are overprovisioned, underused, duplicated or tied to ageing infrastructure and licensing models. Others may have hidden costs in support effort, manual workarounds, specialist skills or operational risk.

A workload assessment should look beyond infrastructure spend alone and consider the full cost of keeping a workload where it is, including the cost of maintaining dependencies, managing incidents and delaying change.

For some workloads, the right answer may be optimisation. For others, migration to AWS or modernisation with managed services may create a stronger case for cost control and long-term scalability.

For telcos, this may mean reducing the cost of supporting complex OSS/BSS, customer care or network operations platforms. For media companies, it may mean lowering the cost of content storage, processing, delivery or adtech operations. For technology firms, it may mean improving cloud economics across product platforms, engineering environments and customer-facing services.

3. Can it support future data and AI needs?

AI-ready workloads need reliable data flows, appropriate access controls, clear ownership and governance that allows data to be used safely.

If a workload holds important, but difficult-to-access data, it can slow down AI and analytics initiatives. It can also increase risk as teams build workarounds to get the data they need.

Leaders should assess whether each workload can support future requirements for automation, personalisation, decision-making and AI-enabled services. Where it cannot, the question becomes whether to optimise, replatform or modernise the workload before AI use cases move further into production.

4. What level of resilience and control does it need?

Not every workload has the same resilience requirements. Some support critical services or customer-facing platforms, while others may be less time-sensitive but still need strong controls because of the data they hold or the processes they support.

A workload-led assessment should consider security, identity, access controls, monitoring, backup, recovery and operational documentation. It should also identify where manual controls or unclear ownership could create risk.

This is particularly important for workloads that support connectivity, streaming availability, subscriber services, customer identity, payments or regulated communications. If teams cannot see how a workload performs, how it fails or how it recovers, it becomes harder to scale AI safely around it.

5. What is the best path forward?

Once leaders have identified value, cost, data needs, resilience and complexity, they can make a more informed decision about the right path for each workload.

That path may be to:

This creates a more targeted route to modernisation. Instead of moving everything at once, leaders can prioritise based on value, risk and readiness.

Turning assessment into a decision framework

These questions give leaders a stronger starting point, but complex TMT estates rarely make workload decisions simple. Applications are often connected through shared infrastructure, data flows, security controls and operational processes. A workload that looks straightforward to migrate may depend on platforms, teams or data sources that make change more complex.

Before committing to migration, optimisation or modernisation, leaders need to validate these assumptions and understand how each decision affects the wider estate.

Here, an Operational Landscape Assessment can help, providing a practical decision framework for modernisation. It can help map workloads by value, cost, risk, complexity and future relevance.

With this visibility, leaders can create a sequenced roadmap to prioritise quick wins, identify high-risk dependencies, plan migration waves and focus investment where it will create the greatest business value.

Rather than forcing modernisation of everything at once, this allows leaders to know which parts of the estate are ready for the next phase of AI-enabled operations and which parts are holding progress back.

Book a migration and modernisation assessment with our team to get a clear picture of your VMware and cloud estate, identify where AWS-enabled modernisation can create value and build the business case for change.