Travelex is replacing multiple legacy core supply chain systems with an off-the-shelf alternative, as it seeks to complete its transition to the cloud.
Meanwhile, as part of its wider transformation, its in-house software development teams work with technologies such as machine learning to develop systems to predict demand and pricing.
The supply chain system replacement project, which is currently at the stage of suppliers bidding for the contract, will begin once one is chosen from the current shortlist of three.
Travelex chief technology officer (CTO) Emanuele Tomeo told Computer Weekly that the company has a digital transformation programme to move services from local datacentres over to the cloud, which has largely been completed. “But certain systems are still in our datacentres because they can’t run in the cloud as they are,” he added.
One example is its core supply chain system. Tomeo described it as the firm’s “backbone”, which enables cash to move efficiently and quickly between its seven global vaults.
“Right now, we have different supply chain systems in regions because it’s old legacy and it’s quite fragmented,” he said. “It’s really hard to maintain and we tried to move it to the cloud, but haven’t been successful in doing so.”
Travelex wants to standardise the supply chain system across its seven regional vaults, said Tomeo. “Essentially we are modernising the supply chain system with one system which will be used across all regions,” he added.
SaaS platform search
The company is currently looking for a software as a service (SaaS) platform as part of a request for proposal (RFP).
“We’re going to become an omnichannel solution for our customers so they can order money in a shop, in a locker at home or from an ATM,” said Tomeo. “In order to be able to do this, and offer as much convenience as possible, we need a modern supply chain system which has capabilities of capturing the order, no matter where it comes from, and then delivering this order where it’s most convenient.”
Work has just started with the requirements being decided. “The approach we are taking is analysing the business, figuring out what’s the simplest and most efficient way of running the vaults and then applying the supply chain system,” he said.
It will probably be about 18 months before the new supply chain software is applied to a first vault.
Tomeo said that once the first is in place, it will become easier to roll out in the remaining six. “We will start from the UK vault, our crown jewels, and then by using the same system that we roll out, we can essentially adapt it for another regional vault,” he added.
In-house development focus
There is a great deal of activity in the IT department at Travelex. In 2020, it entered administration, after the Covid-19 pandemic and an attack by cyber criminals caused major losses.
A combination of declining demand as the travel sector slowed during Covid-19 restrictions and the downtime and losses that resulted after the Sodinokibi cyber crime group ransomware attack put the company on the brink.
Today, with new investors, it continues to build for its future, with digital technology at the core.
Tomeo joined as CTO, a position that was new, about a year ago. “I essentially took over what Hans van der Waal, director of global IT, was doing in terms of the IT department and the product development side,” he said.
Van der Waal remains, but Tomeo took charge of software engineering teams and technical personnel. “Let’s say product-led and product-focused rather than execution,” he added.
While the company is moving to commercially available software for its supply chain backbone, it remains busy developing its own. Travelex has a team of 400 IT professionals out of an overall global workforce of about 5,000, with two-thirds based in Mumbai in India.
“Software for things like stock forecasting or price predictions will be built in-house,” said Tomeo. “This is why we’re investing also in machine learning and AI [artificial intelligence], in order to be able to predict, let’s say, stock levels that we will need.
“We are working now on some machine learning models to be able to predict based on both seasonality and past trends, and also interacting with current data like flight and major event information,” he said.
Its internal teams are harnessing GenAI with a pilot that includes almost half of the developers, said Tomeo. “We are equipping them with the tools, but also best practices and prompts, measuring the productivity gains in different areas and use cases,” he added.
“Currently, we’re seeing the major gains in writing documentation, test generation, code explanation, some code generation use cases, and data clean-up and operations,” said Tomeo.
He said in some areas there have been productivity gains of between 20% and 40%. “The current conclusion is that, if used well, it can boost developer productivity, but it takes time to understand how to best use it,” added Tomeo.
Travelex is also investing in a modern data platform, which he said will be the foundation for all analytics and insights, as well as supporting global real-time AI products. The latter includes predictive analysis of customer behaviour, forecasting of stock levels; and dynamic pricing, both of which the company is currently experimenting with.
“This is particularly challenging given the number of fragmented systems and global presence, so centralising into one place will massively improve our effectiveness and speed to market,” said Tomeo.
