Hays

Deadlines compress decision making. They remove bureaucracy and force teams to focus on what actually matters. In Hays’ case, it created a rare opportunity to leave behind more than a decade of accumulated technical debt, duplicated services and technology choices that made sense once but were now costing efficiency. 

By moving with discipline and structure, the team managed to simplify the estate, adopt a cloud only direction and shift several services into environments where they would be easier to maintain, scale and secure. They did not attempt a wholesale transformation of every application in the estate. Instead, they brought sensible modernisation forward wherever it would create immediate and lasting benefit. 

This combination of necessity and clarity is what turned a high-risk migration into a strategic advantage.

By the time the first workloads were ready for migration, something important had happened. The team had shifted away from treating every migration as a standalone effort and had instead created a library of repeatable patterns. These patterns defined how a category of workload should be assessed, built, secured, tested and cut over. This pattern-based approach allowed for automation, minimising manual tasks and increasing the cadence and accuracy of every migration and cutover. 

Once those patterns had been validated in dev and test environments, they were used across hundreds of workloads. As Simon described it, “if it works in test on that server, we can roll it out across these 400”. That level of consistency made delivery faster and it also reduced risk because the team knew that each workload was following the same proven approach. 

The effect of this was significant. Confidence increased with every successful wave. The challenges that looked large in the early weeks became more manageable as the programme moved forward. Progress accelerated not because the team worked harder, but because the process became cleaner and more predictable.

Despite the progress, there was a moment where the entire programme could have taken a different turn. During one of the late evening cutovers, the team identified a technical issue that was not catastrophic but was unusual enough to create concern. They were already in the change window. Everyone was present. The natural instinct in many organisations would be to push through and deal with the consequences later. 

Instead, the engineer leading the cutover made a call that changed everything. They simply said, “we are not going to go live tonight”. It was a moment of clarity, judgment and maturity. 

Over the next three days, the team diagnosed the issue properly, resolved it and completed the migration without any impact to the business. There was no downtime, no noisy incident afterwards and no lingering problems waiting to appear weeks later. It was a demonstration of why empowered engineers, psychological safety and disciplined patterns matter as much as technology.

One of the most transformative outcomes of the migration was the improvement in operational resilience. Before the move, disaster recovery testing required significant planning and often relied on extended weekend windows. After the migration, the estate was modernised to the point where component level DR testing could be performed during normal working hours. As Simon said, with clear relief, “we have our Easters back”. 

This shift captures the true value of cloud. It is not the location of resources that creates advantage but the ability to automate, scale, simplify and test without causing disruption. The estate became easier to manage and the teams became more confident in their ability to respond to issues quickly. 

Hays now has a foundation that allows them to adopt AI-based operational improvements in a meaningful way. Rather than talking about AI abstractly, Simon described it as an intelligent assistant that can help teams spot patterns, identify anomalies, prevent capacity issues and strengthen the security posture of the estate. 

This is only possible because the environment is now sufficiently standardised, observable and well governed. In other words, the modernisation created the conditions for future innovation rather than trying to force innovation onto an unstable base.