I’d pick a few examples. The first one is quite recent. We started building our CDP using one of the classical solutions available on the market. But the problem with most of these platforms is that they’re designed for companies focused on handling second or third-party data.

Accor has a lot of first-party data. So, we stopped and launched a new RFP to find a better-fit CDP. That’s when the team proposed something quite innovative: A composable CDP. Unlike the classical architecture where you have to move your data into the CDP and adapt to its model, a composable CDP sits on top of your existing data warehouse or enterprise data lake. It adapts to the data you already have. You just need to build the bridge between your business vocabulary and the way the data is stored — in our case, in Snowflake.

This completely changed the game for us. In just a few months, we had the platform ready, introduced it to internal users, trained people, and put it into action. We solved the challenge in a short time, thanks to a team that wasn’t afraid to make bold choices and think beyond market standards.

Another example is our GenAI-powered travel assistant. We were one of the first hotel players to launch a beta test of a travel assistant using a large language model (LLM) to book trips and hotels on our website. We ran the pilot with specific user segments in Australia and England.

The results weren’t good. One of our key findings was that we hadn’t thought enough about how to integrate these new tools into the customer journey. It’s not just about launching a new feature, it’s about making sure it’s part of the flow and used at the right moment.

For example, if you’re going to Berlin next weekend with your wife, you probably don’t need a travel assistant. You’ll get there faster using a standard filter-based search. But if you don’t know where you want to go — maybe you’re looking for a sunny beach, a volcanic area, a forest, or something peaceful within three hours of Paris — that’s where GenAI can really shine. That’s when the assistant becomes helpful. Making sure we direct guests to the right tool at the right moment is something we had overlooked, and we’re adjusting to that now.

We also need to think more about the ROI of GenAI. It’s not just a financial cost, it also has an environmental impact in terms of CO2 consumption.

For a large enterprise like Accor, deciding where to place the data organization is a complex decision. Some organizations position data closer to IT, which is great because you have skilled people, and for the data team, it can be quite exciting — you get to build your expertise, and so on. But the risk is becoming disconnected from business understanding.

It’s critical for the data team to stay close to the business because you need to understand the meaning and value behind the data. On the other hand, if you position the data team too close to the business, you may have a great time to market at first and deliver value quickly, but you might overlook the importance of industrialization: building on solid foundations, investing in technology, and so on. At the end of the day, data is still very much an IT topic.

We’ve done something really effective — we’ve placed the data team within an organization that includes both the most critical digital business units of Accor and the IT team.

So we are right in between. That positioning has unlocked a lot of potential and supported our domain-oriented approach. Our team is fully dedicated to one business domain, understanding it deeply while also staying very close to IT. That allows us to take the time to build strong foundations, invest in technology, and more. Ensuring the data organization is not too far from the business and not too far from IT is key to achieving success with data.

It’s a very complex question because there are many ways to come to data. I benefitted a lot from my career starting on the field. First, as a developer, doing a lot of different aspects of development. At the very beginning, when I was in a startup, I was in charge of the server, to be racked, making sure that the disks were working effectively.

Then I moved from development to backend development, then to architecture. Another turning point in my career was when I decided to specialize in performance optimization: making sure that software performed correctly, optimizing the disk, the CPU, the memory, and solving crashes and all that. That changed the way I was developing and thinking about architecture.

It helps a lot when you go into data because data is a lot about thinking smart in terms of architecture. You can easily run into difficulties, especially since data comes with many performance challenges. It’s important not to move too fast and understanding the complexity behind it really helps.

The second point is that if you want to go far working in the data world, you need to build your business acumen. You have to focus not just on doing exciting things with the latest technologies but also on adapting to your business problems — bringing value and staying close to your business needs. That is key for modern data teams.