How FCM Travel went from "AI access" to "real adoption" with Cambridge Spark

About FCM Travel

Flight Centre Travel Group (FCM Travel / Corporate Traveller) is a global corporate travel management group with over 15,000 employees. Its technology hub in Spain, with around 115 people in Barcelona, is the centre where this Data and AI training programme was concentrated.

The challenge: from "human-first" to "AI-first"

FCM Travel came with a clear mandate from management - CEO, global CTO and the board - to transform their workflows from "human-first" to "AI-first". In parallel, the company had signed an agreement with Anthropic (200 licences) and initiated the migration to Claude Code.

The challenge was not access to the tools, but closing the skills gap homogeneously across the technical organisation: standardising training, establishing good prompt engineering practices and familiarising teams with protocols like MCP. Added to this was an uneven cultural adoption, with very enthusiastic profiles co-existing with more sceptical ones and legitimate security concerns.

Without a common foundation, the risk was to progress in a fragmented way, with each team learning on their own, without a shared language or practices, which would have slowed down real adoption and made it difficult to measure its impact.

At the FCTG Tech Hub in Barcelona, AI transformation is our top priority. More than technical training, we were looking for a mindset shift: for our team to embrace transformation across all phases of design, deployment and discipline. Cambridge Spark was a great partner for this challenge. We worked hand in hand with their team to define our needs and build a bespoke solution. Thanks to this collaboration, we have ensured the whole team shares the same language and knowledge, and is prepared to drive our AI journey forward.

At the FCTG Tech Hub in Barcelona, AI transformation is our top priority. More than technical training, we were looking for a mindset shift: for our team to embrace transformation across all phases of design, deployment and discipline. Cambridge Spark was a great partner for this challenge. We worked hand in hand with their team to define our needs and build a bespoke solution. Thanks to this collaboration, we have ensured the whole team shares the same language and knowledge, and is prepared to drive our AI journey forward.

The solution: differentiated pathways by profile

Cambridge Spark designed training pathways tailored to each profile, instead of a single generic content for everyone:

• LLM Engineer Pathway

• AI Product Manager

The programme was coordinated from the FCTG Tech Hub in Barcelona, with Guillermo Ortiz Figueroa leading the engineering pathway and Sarah Petit leading the product one, in parallel.

Beyond delivering the pathways, what FCM Travel valued most was Cambridge Spark's willingness to iterate: after the first cohort, internal feedback was gathered and Cambridge Spark redesigned the following module with greater technical depth, a change of format to half-days, a preliminary level assessment and more practical exercises.

We needed that initial push at an organisational level, the first module was precisely what triggered the right internal conversations. Today we have massive and in-depth AI adoption across engineering and product, and that first step with Cambridge Spark was the starting point.

We needed that initial push at an organisational level, the first module was precisely what triggered the right internal conversations. Today we have massive and in-depth AI adoption across engineering and product, and that first step with Cambridge Spark was the starting point.

The results: the "push" that triggered a complete transformation

The greatest impact of module 1 was not in the initial survey, but in what it triggered at an organisational level: it was just the push that the organisation needed. Beyond the numbers, what really made the difference was the effect it had within the company, generating a momentum that far exceeded any single assessment.

That first module generated the internal conversations that accelerated everything that followed: the agreement with Anthropic, the deployment of Claude Code for the entire engineering team (and Claude Desktop for key profiles), and a complete transformation of the software development lifecycle.

Adoption today is massive. In just 3-6 months, FCM Travel went from virtually zero consumption to intensive daily use, with its most active engineers generating a volume of usage that today is, in fact, a cost management challenge in itself — a clear sign of the depth of real adoption, not just nominal access to the tool.

Beyond the figures, the training was the catalyst for a fundamental cultural shift: FCM Travel went from a small group of early adopters to widespread adoption in the Barcelona hub and its distributed teams. This transformation extended to the entire software lifecycle — from QA towards SDET — and is also influencing how new products are conceived, incorporating AI natively where it makes sense.

Looking to the future

Today, in engineering, Claude Code is already the main daily development tool, integrated. throughout the software lifecycle: from the definition of requirements (Spec-Driven Development) to code review and testing. In product, AI is used as an active partner in the design of new features for travel agents.

FCM Travel and Cambridge Spark have already agreed on the next steps: to continue deepening technical knowledge by level, to maintain the format adjusted after the first cohort (more depth, prior assessment, half days) and to extend training to the distributed teams that are now adopting these practices.

A very interesting programme. It was exactly what I needed to go beyond daily use and delve into how LLMs actually work. The sessions on tensors and LLM architecture were the highlight for me: they provided me with the missing mental model and a solid foundation on which to build. And most importantly, they sparked my curiosity to continue delving into the subject, which is what I expected to get out of this course.

A very interesting programme. It was exactly what I needed to go beyond daily use and delve into how LLMs actually work. The sessions on tensors and LLM architecture were the highlight for me: they provided me with the missing mental model and a solid foundation on which to build. And most importantly, they sparked my curiosity to continue delving into the subject, which is what I expected to get out of this course.