The CEO Corner

Emerging tech and AI

Which emerging tech trends are you most excited about?

Aiman Ezzat: Generative AI [Gen AI] and agentic AI are at the forefront. AI is evolving from copilots to autonomous, reasoning agents capable of orchestrating complex tasks across supply chains, predictive maintenance, and customer engagement. This shift will enable new AI ecosystems that drive efficiency, resilience, and innovation at scale.

Second, large language models (LLMs) are transforming robotic capabilities to near-human levels and surpassing human capabilities in some tasks. Next-gen robots can handle complex, interconnected tasks, improve decision-making, and enhance operational efficiency across industries.

Third, there’s growing interest in sustainability-focused innovation. AI is not only transforming industries but also driving the resurgence of clean energy sources, such as nuclear, to meet the growing demands of digital technologies.

Where do you see AI and, in particular, Gen AI having the biggest impact?

Aiman Ezzat: Agentic AI represents a new frontier of digital transformation, accelerating efficiency and value creation by automating highly variable processes that traditional technologies cannot automate. I believe AI agents can transform the way in which businesses operate. Successful enterprise operations require AI agents working seamlessly with humans, under human control. I like to call this human-AI chemistry. We see Gen AI assistants and AI agents as tools to assist people (such as coding assistants) by augmenting and amplifying human ingenuity (for example, in developing new molecules). Most professions will be affected by AI one way or another and, whatever their seniority, they should learn how to use these tools. This should be a massive change-management priority for business leaders.

In the short term, the most promising fields are business process operations with agentic AI workers for hyper-automated operations 24/7; go-to-market, with enhanced customer targeting, experiences, and interactions; IT, as in software development with coding assistants and agents; testing automation; knowledge management, such as creating documentation, etc.; and in the delivery of greater value in terms of CX and new products delivered faster and better, including with improved R&D (such as developing new molecules or lighter parts in automotive/aerospace).

Christel Heydemann: AI has already had a huge impact at consumer level. Interestingly, a lot of that usage today is for fairly simple things, like getting answers or summaries quickly.

In the professional world, one of the biggest impacts we see right now is in software development. AI coding assistants are speeding up programming, and that affects many tech companies. At Orange, it’s also changing how we work by bringing a lot of efficiency in-house. AI makes a difference in almost every routine process, from automating fraud detection in banking to streamlining compliance checks. In the telecom sector specifically, we’re working on what you might call AI-driven or “intent-driven” networks. We’re also applying AI to improve customer service. Gen AI offers more personalized self-service tools, smarter chatbots, and so on, which ultimately raise customer satisfaction.

I also see great potential for AI to be used as a personal coach or assistant for employees . Think of an AI mentor that helps you learn new skills or navigate challenges at work, at your own pace. Today, if you want to upskill, you might take a course or ask a manager – and sometimes people feel shy or judged in those situations. But an AI tool is completely non-judgmental. It’s just there to help you whenever you need it. I believe using AI in that way – to empower employees in their own development – could be very impactful.

---

Conversations for Tomorrow #10: Crafting technology

The tenth edition of u003cemu003eConversations for Tomorrow u003c/emu003edelves into key technology trends that are impacting organizations and society in 2025 and beyond.

Read more

---

What do you think is the biggest hurdle to scaling AI?

Aiman Ezzat: It’s not the technology. It’s trust, governance, and organizational readiness.

While most organizations have moved beyond pilots, only 9% have successfully deployed Gen AI across multiple functions/locations. This highlights the challenge in operationalization, rather than experimentation. Capgemini research shows that the portion of organizations that trust in fully autonomous AI agents has declined from 43% to 27% in the past year , and many have yet to put in place robust governance frameworks. This lack of trust and governance slows adoption and creates risk.

Finally, human-AI collaboration and cultural transformation are often underestimated. In the short to medium term organizations will have AI agents working within human teams. Blended teams – where humans and AI agents collaborate – will become the norm, driving productivity and innovation. Reskilling, new operating models, and fostering “human-AI chemistry” are essential to sustainable, trusted AI adoption. In short, scaling AI is less about algorithms and more about building trust, governance, and a culture ready for hybrid human-AI teams.

Christel Heydemann: Anyone can play with ChatGPT and get it to draft an email or summarize a document. Those use cases are relatively straightforward and can bring quick wins. And as long as you put data safeguards in place, rolling out those basic tools to employees isn’t too difficult.

But if you want AI to fundamentally reinvent processes, that’s a much bigger challenge. You need well-organized, accessible data. You might have to re-engineer processes to integrate AI effectively. Often, you even have to change the culture of the organization, so that people trust the AI enough to use it in their daily decisions. Scaling AI isn’t just a matter of installing some software. It means rethinking entire workflows that have been in place for year. That kind of change doesn’t happen overnight.

Telecoms and AI

What impact do you think AI will have on telecom networks, and what does an AI-driven network look like?

“We anticipate that, by 2030, around two-thirds of all network traffic could be related to AI in some way”

Christel Heydemann

AI-driven traffic will have different patterns and requirements. If you look at the history of networks, 20–30 years ago, most traffic was voice calls. Then the internet brought a surge in data traffic. More recently, video streaming exploded, which meant our networks had to handle a ton of data going from the network to users. Telecom operators responded by building content delivery networks and beefing up capacity to make video streaming smooth. Now, with AI, we expect a lot more upstream traffic (devices and sensors sending data to the cloud for processing, or users uploading content for AI to analyze). Some AI applications will also demand low latency. For instance, if you have an AI-driven control system in a factory, it needs real-time responsiveness.

Our current network architectures are designed mostly for heavy downloading (video, web browsing, etc.). We’ll likely need to re-engineer parts of the network to cater to this new AI-driven pattern.

On the operations side, AI is going to be indispensable for managing and securing complex networks. Keeping everything running optimally is a huge task. AI can help by analyzing vast amounts of network data in real time and adjusting or flagging issues. For example, an AI system could predict that a certain network node will become congested in the next five minutes and proactively reroute traffic, or it could detect a hardware fault and alert us to fix it before it causes an outage. AI can also filter out the noise. In network operations centers, we get thousands of alerts, and many are false alarms. AI can learn to tell the difference and only raise the truly important issues.

In cybersecurity, AI is a tool for both attackers and defenders. We’re using AI to strengthen our defenses: detecting unusual patterns of network traffic that might indicate an attack or identifying malware. But we know attackers are also using AI to find new vulnerabilities or to automate attacks. So, it becomes a bit of an arms race. An AI-driven network will likely include AI “guardians” that continuously monitor and protect it. Manual monitoring just can’t keep up with the speed of attacks nowadays.

---

~2/3

of all network traffic could be related to AI in some way by 2030, according (C. Heydemann)

~9%

of organizations have successfully deployed Gen AI across multiple functions/locations (A. Ezzat)

27%

of organizations trust in fully autonomous AI agents, down from 43% in the past year (A. Ezzat)

---

Aiman Ezzat: AI will fundamentally transform telecom networks by making them autonomous, predictive, and service-centric. Tomorrow’s AI-driven networks will be self-configuring, self-healing, and self-optimizing, enabling telcos to deliver superior performance and new revenue streams.

AI will integrate deeply into network operations, enabling real-time anomaly detection, predictive maintenance, and proactive issue resolution. Gen AI is already revolutionizing network operations by providing intelligent document querying, automated troubleshooting, and conversational interfaces for technicians, reducing downtime and improving resilience. Agentic AI will unify customer service and network operations, creating end-to-end automation that eliminates silos and accelerates problem resolution, while reducing costs and penalties.

How do you see the future of the telecom sector, and what role will software play?

Christel Heydemann: The telecom sector is at a crossroads. On one hand, telecom has been the backbone of huge technological shifts – for example, the shift to mass cellphone use. Our networks were (and are) the foundation that made innovations like smartphones useful. Most of the time, though, people don’t even think about the network until something goes wrong with it. Then, we realize how central it is to modern life. Also, if there’s a natural disaster (floods, wildfires, storms) or a major cyber incident, people rely on telecom networks to contact loved ones and get help. It’s a responsibility we take very seriously.

This ties into the question of scale. In Europe we have a very competitive telecom market but, currently, it’s too fragmented. For instance, the US and China each have just a handful of large telcos, whereas in each European country you might have three, four, five operators fighting over a relatively small market. This makes it hard to achieve the scale to invest heavily in new technologies and get strong bargaining power with equipment suppliers and phone manufacturers. The European regulatory environment was created about 20 years ago, when the priority was breaking up monopolies and encouraging competition within each country. Back then, that made sense. It brought prices down for consumers and drove adoption. But, despite the market having matured, in 2025 we’re operating under largely the same rules. We need to rethink those rules to reflect new realities.

---

Telecoms: What’s next?

At Capgemini, we help you reimagine the future, enabling telco organizations and network equipment providers to deliver a profound impact on their business, their customers, and the world at large. 

How we can help

---

Recently, an EU-commissioned report by [former European Central Bank President] Mario Draghi made this point very clearly: Europe should modernize its telecom regulations and encourage consolidation . I agree with that perspective. Growth is about new services, rather than new customers. In a low-growth environment, having five competitors where two or three would suffice means revenues get spread too thinly. Telecom is a business with high fixed costs. You need a certain scale to be efficient. If we can achieve greater scale (for example, through mergers or partnerships), we can be more efficient and invest more in innovation and network upgrades.

Scale also matters when dealing with global tech giants. A bigger, consolidated European operator would have more sway when negotiating with Apple or Google on things like network features or revenue-sharing models. Right now, if you’re a smaller operator, it’s hard to influence those discussions.

The networks of the future will still be our core, but we’ll develop more software, more services, and more partnerships to ensure we capture a fair share of the value. Providing excellent connectivity is non-negotiable. But we also want to grow beyond connectivity, so we remain competitive and relevant in the digital economy.

You’ve mentioned Europe and sovereignty. Why do you believe consolidation is so important in this context?

Christel Heydemann: Technological sovereignty is about a region or country maintaining control over its tech destiny. This concern comes up a lot in Europe, as well as Africa and the Middle East. Leaders everywhere are asking: How do we make sure we don’t lose control over our future? Europe has lost ground in many areas of tech over recent decades (there’s no European equivalent to Google or Apple), so we have to be smart and defend the strong positions we do have, while building new capabilities for the future.

One of Europe’s challenges is that, on a global scale, even our biggest companies are relatively small. In telecom, for instance, Orange is the second-largest operator in Europe, which sounds great, but in global terms our size is modest. When we negotiate with giant companies like Apple, Google, or Amazon, a single European operator – or even a single European country – doesn’t have a lot of clout. These companies operate across the entire world, with billions of users, meaning they set terms. That’s a tough position for us, as European businesses, to be in.

This is where consolidation comes into play. By consolidation, I mean encouraging the formation of larger, stronger European entities (through mergers, alliances, etc.) that can stand toe-to-toe with the global giants. If we had, say, a pan-European telco, instead of many smaller national ones, that larger entity could invest more in new technology, achieve better economies of scale, and have more influence in partnerships or negotiations. The same logic can apply in other tech industries. Size isn’t everything, but it does matter when you’re competing globally.

The EU has recognized these issues. We’ve seen new regulations like the Digital Markets Act (DMA) and Digital Services Act (DSA) introduced to curb the dominance of the big global tech platforms. The upcoming AI Act is designed to ensure AI in Europe respects our values. All of these are important steps to protect consumers and competition. However, regulation alone is not enough to guarantee Europe’s tech sovereignty. We also need our own champions. That was a key message in the report led by Draghi: Europe must build competitive European firms. Instead of 10 fragmented markets, act as one big market, so our companies can scale up.

Sovereignty isn’t just about who runs the networks. It’s also about who develops and controls AI and the other new technologies that will shape the future. We have amazing car companies, energy companies, pharmaceutical companies in Europe. If they all embed AI and become more competitive, Europe stays strong. If they hesitate because of lack of resources or fear of the unknown, we risk falling behind not just in tech, but in those industries, too.

We should regulate AI to address risks (privacy, bias, etc.), but we must not over-regulate to the point that we stifle innovation, because other regions will absolutely forge ahead. Europe should aim to lead in areas such as ethical AI, industrial AI, AI at the edge (where we have some advantages with our engineering and manufacturing base). We don’t have the dominant social media platforms or the dominant smartphone operating systems, but new battles are coming with AI and other tech. We have to position ourselves to win some of those.

How do you use AI in your day-to-day role as a CEO?

Aiman Ezzat: I believe Gen AI can advance beyond simply being a productivity tool to become a co-thinker for managers in organizations of all sizes, aiding in problem-solving and decision-making. It can become a sparring partner, offering fresh perspectives and challenging assumptions, even enhancing strategic thinking and leadership development. It is one of the inputs I take before I make a decision.

Christel Heydemann: To be honest, I don’t use AI in a very heavy or specialized way day-to-day – at least, not yet. But I do take advantage of some of the tools out there. For example, instead of doing a traditional Google search for information, I might put a question to our internal AI assistant or to a tool like ChatGPT to get a quick, synthesized answer. It often gives me a more concise answer than wading through pages of search results.

The biggest boost I get from AI is in managing information flow. I receive a lot of lengthy documents – reports, presentations, analysis. We have an internal AI-powered tool (it’s like our own ChatGPT trained on corporate content) that can generate summaries of these documents. So, if someone sends me a 50-page PowerPoint deck, the AI can produce an executive summary or bullet-point highlights. This has been a game-changer in terms of saving time.

Of course, as a CEO, if it’s a critical matter I won’t rely solely on the AI summary. But it’s a fantastic starting point. If something in the summary catches my eye, I’ll jump into that section of the full document to learn more. It’s a way of triaging information.

---

Rise of agentic AI: How trust is the key to human-AI collaboration

AI agents are becoming a core part of enterprise operations, reshaping business models, workforce dynamics, and competitive advantage.

Read our recent report

---

I’ve also been working to train and educate myself and my executive team on using these tools effectively. We held a workshop where we experimented with writing prompts to get better outputs from AI. We’re extending that kind of training throughout the company. This includes understanding their limitations. I enjoy testing the AI and seeing where it might misfire. I’ll sometimes deliberately ask it a question to which I know the answer, just to gauge its accuracy. It’s a reminder that you still need a critical eye. AI can streamline a lot of things, but you can’t 100% outsource your judgment to it.

Another area where AI comes into my work routine is communication. I know some of my team members use AI to draft emails and even brainstorm ideas for presentations. I’ve tried out Gen AI to see if it can rephrase a complex message more clearly, or to summarize long email threads. I’m not a salesperson, so I’m not using AI to, say, write sales pitches every day, but I know our sales and marketing folks are exploring those applications. For me, it’s more about efficiency in reading and research. And I do find it helpful. AI shaves hours off of mundane work for me each week, which is pretty significant. It lets me spend more time on the human aspects of the job that AI can’t do, like strategy, mentoring, and so forth.

Women in leadership

---

How can we break down the barriers to more women reaching senior leadership roles?

Christel Heydemann: That’s a multifaceted challenge that many companies and societies are grappling with. I think the solution lies partly in empowering women themselves, and partly in changing the culture around them.

We often find that women, even extremely talented ones, are less inclined to put themselves forward for promotions or high-profile projects . There’s a lot of research showing that a woman might feel she needs to meet 100% of the job criteria before applying, whereas a man might only feel he needs to meet 60%. So, one barrier is this confidence gap and a tendency to self-withdraw from opportunities. We can encourage women through mentorship and sponsorship programs, and by creating an environment where they feel their contributions are valued. I always say, the more we showcase strong female role models, the more other women will see a path for themselves . It’s inspiring to see someone who has a similar background or faced similar challenges succeed. It makes it seem more achievable for the next person.

On the company side, it starts with a genuine commitment at leadership level. At Orange, we’ve made it a clear priority to help women progress. That means we set measurable targets (for instance, increasing the percentage of women in top management by a certain amount), and we track progress as we would with our financial metrics. We also have initiatives such as leadership development programs specifically for high-potential women, and training for all employees to address unconscious bias. These efforts signal that we’re serious about change. And it requires sustained effort. It’s about building an inclusive culture where women have equal access to opportunities and feel valued, personally and professionally.

Aiman Ezzat: Breaking down barriers for women to reach senior leadership roles starts with disrupting persistent gender stereotypes . Our research found that, while the vast majority of leaders agree women are as effective as men, stereotypes around future-critical skills such as AI, automation, and data analysis remain deeply entrenched. Nearly half of male executives in our research perceive these technical skills as “masculine.” Left unaddressed, this bias could widen the leadership gap.

There’s a few things leaders can and should do:

• Interrupt bias systematically: Train leaders to spot and challenge bias and make hiring and promotion criteria transparent.
• Embed technical fluency: Ensure all leaders, regardless of gender, have access to training in AI and data skills, which are increasingly vital to advancement.
• Democratize sponsorship and mentorship: Provide equal access to mentors, sponsors, and high-visibility assignments.
• Normalize flexibility: Make flexible work options available to everyone, supporting both women and men in balancing work and life.