• While AI enhances the efficiency of senior staff working on complex architectures, it poses a challenge for the integration of junior staff. It is also driving Orange to strengthen its internal expertise.
• Frédérique Hachmi, Director of IT for the DATA-AI domain at Orange France, explains that the success of an AI transition depends on curiosity, a leadership mindset, and close collaboration with business units.
How do you support Orange developers in adopting artificial intelligence?
We have rolled out tools to help and assist developers, with the goal of supporting them in their day-to-day work and exploring the practical applications of AI. Beyond simple code generation, we’re working on key features: unit tests, regression tests, and documentation, among others. We aim to foster a positive momentum and encourage initiatives, for example through “AI challenges” within IT teams. The goal is to help them automate tedious production tasks.
We are increasingly looking for ‘conductors’ who can take a step back and oversee the integration of agents, data, and tests.
What are the benefits of AI for improving code documentation and testing?
Historically, documentation has been a time-consuming task that is not a high priority for developers. Yet documentation is vital for the long-term maintenance of applications and infrastructure, especially since some of them are over ten years old. AI now makes it possible to better structure documentation in real time, which facilitates the subsequent takeover of the code by other team members. The same is true for testing: because we always have to move quickly, we only cover a limited portion of the code to be tested. Now, AI enables us to deliver applications with a high level of availability and service.
What skills are expected of a developer today?
Technological curiosity is at the heart of our expectations: they must demonstrate an appetite for tools like GitHub Copilot or Claude, and be capable of staying constantly up to date.
Developers will largely rely on the same skills but apply them differently. The real change lies in the hybridization of roles. We are moving toward technical architects who can take a step back to oversee the integration of agents, data, and testing while ensuring the security of solutions.
Among the criticisms we see regarding the widespread adoption of AI among developers, the topic of a future imbalance between junior and senior developers in teams often comes up…
This is indeed an issue to watch, as senior developers use AI so intensively that they sometimes max out their request quotas. And paradoxically, the market is tightening for junior developers. This is largely because AI can easily generate modern code like Python, which can put junior developers at risk.
However, AI performs less well with backend structures or older languages, where human expertise remains essential.
This is therefore pushing developers to specialize in areas where human expertise remains indispensable. Our challenge is to better support junior developers in this context, particularly through mentoring pairs with senior developers, because we will eventually need to be able to replace our current leaders.
Should most tech companies expect to reduce their workforce?
Companies like Anthropic cite an efficiency factor of 1 to 5 thanks to AI. At Orange, this will impact our production model: it is highly likely that our use of external service providers will shift in terms of the nature of the work and decrease. In other words, we want to refocus our internal talent on orchestrating agents and end-to-end integration within our information system. The use of external resources will be limited to very specific needs.
Will companies have to undergo a major reorganization?
An organization is a living entity that constantly adapts in an environment where technologies evolve rapidly—very rapidly. Unleashing creativity and fostering agility are the keys to this resilience, just as much as employee engagement.
AI is transforming the way we work: we’re collaborating more closely with business units, using increasingly shorter iteration cycles to address their challenges and deliver the expected value. It also requires greater flexibility—seeking out expertise wherever it exists, beyond the usual silos. A developer with expertise in networking can step in to help with a consumer-facing issue overnight. It is this agility and adaptability—beyond technical skills—that will now define what it means to be a developer in the era of ethical and responsible AI.
This text has been translated by an artificial intelligence.
Frédérique Hachmi
