Bureau Veritas SA (BVI) Earnings Call Transcript & Summary

Thank you for the questions, [indiscernible]. Good questions. Of course, it's a defining move, right? It's such a large competitor. I think what it means is, in fact, -- that move reinforces our conviction that we need to advance and progress on our multi-specialist. Growth is going to require that focus. Our approach to focus, we never really talked about multi-specialist when we launched in 2024, but I can tell you that was the jargon we were thinking about and really the direction we wanted to go to because the market uptick of testing, inspection and certification has evolved and requires a lot more technology. It requires a lot more new ways of working that we didn't have in the sector before. So we were working in that direction. And the fact to see one of the competitors in that position, it reinforces our conviction is the best way to grow is to do all these things so we can become a more valuable company. So it's a move that to us reinforces our strategic direction. That answers your question. I don't know if anyone wants to comment.

I have just another comment perhaps. I think it's a great move. It shows investors have an interest in this segment. And I think some of you have been following Veritas for quite long. And we've gone through -- the sector has gone through some years of less interest. I think those years have definitely gone. We've launched here at Bureau Veritas, the first movement of starting to dispose some of the portfolios a couple of years back with movements happening. Now this movement is happening with Intertek. I think it's quite interesting. It shows that there is interest and then some of the valuation here around being questioned. So all in all, the growth agenda that Hinda mentioned is the most powerful tool for Veritas to demonstrate to create value. You consider the Intertek story, obviously, they were the least growing entities among the listed assets.

Okay. So thank you all for your questions. We'll now have a 20-minute deserved break, coffee break. Feel free to ask questions during the break. And let's be back at 11:10 -- we'll restart at 11:10 sharp. So let's be back at 11:05. [ Break ]

We will be focusing on 2 key pillars of our LEAP I 28 strategy, performance on one side and people on the other side. So first, Noor Sait will take us through how we are strengthening performance across the group. We then turn to our people agenda with Maria Fraguas Lorente and because no discussion about the future will be complete without innovation, Philipp Karmires, our Chief Digital & Innovation Officer, will join us to discuss how AI is already transforming the way we work, serve our clients and create value. So please join me in welcoming Noor to the stage.

Good morning, and welcome to the Performance-Led Execution presentation. My name is Noor Sait, and I'm the Chief Performance Officer at Bureau Veritas. I joined Bureau Veritas a year ago after spending more than 30 years in the upstream oil and gas industry. During that time, I led and grew engineering, manufacturing and field operations businesses in several locations around the world. Improving performance has been central to all my roles, and I'm very excited to bring this experience to Bureau Veritas and to join an impressive ongoing effort to improve how we work at scale. Over the next few minutes, I will walk you through the performance-led pillar. I will share what has been delivered at the midpoint of the LEAP I 28 strategy against our commitments that we made in 2024. And I will show you that the performance agenda is a structured effort with tangible progress and results applied across a large portion of the organization. It consists of multiple programs and projects designed to improve how we operate and how we work and how we deliver value. I will begin by reconnecting the program to our stakeholder commitments. As Hinda has shared earlier, our purpose anchors our plans and vision and defines the commitments we make to all our stakeholders. For our clients, that means putting them at the center of what we do and the experience we provide them and the novel solutions we develop, all supported by the LEAP I 28 strategy and especially by the portfolio pillar that Maria has just covered. For our employees, it means helping them progress and develop to bring growth and impact, and this is covered in the evolving -- in the Evolved people model, which Maria will present later. For society, it means operating transparently and contributing to a more sustainable world as reflected in the CSR progress I will share with you. And for our investors, it means stepping up our operational performance to unlock growth, improve margins and create the capacity to reinvest in the business. Performance-led execution, therefore, sits at the heart of our stakeholder commitments. It is not just a financial imperative. It is how we build and sustain trust with every stakeholder that we serve. Turning to our corporate social responsibility commitments. Our CSR plan is built around our vision to shape a better world through a better environment, a better workplace and through better business practices. This journey started with the Bureau Veritas sustainability framework and continues through engagement with external stakeholders, our Board of Governance Committee and our employees around the world. Our targets cover 5 priorities, and we are making good progress across all of them. This progress is recognized through our external ratings such as our #1 ranking from Sustainalytics, our #2 sector ranking by S&P Global and our gold rating from EcoVadis. I will now move to the performance programs that support our investor and customer commitments. So performance and execution consists of 2 streams. The first is operational leverage and functional scalability and the second is innovation. Operational leverage and functional scalability focus on improving efficiency and productivity, reducing our cost to serve, differentiating through superior customer service and supporting growth. The second stream, innovation, provides the digital backbone that enables our performance programs and supports new services and delivery models. Philipp will take us -- tell us more about that in the digital and AI session next. My presentation focuses on operational leverage and functional scalability and the innovation embedded in the performance programs that I will share. As we shared in 2024, performance-led execution has one clear financial objective. It is to deliver 180 basis points of cumulative margin improvement from 2023 to '28 with half of the gains reinvested to fund modernization and growth. We will now look more closely at the operational leverage and functional scalability streams. The slide summarizes how these streams are organized and the programs and the projects under them as well as the commitments and progress. Under the operational leverage stream, we have 3 programs supported by multiple projects with an ambition to deliver up to 100 basis points of cumulative margin improvement from '23 to '28. The 3 programs under the stream are performance management, process improvement and tech augmented ways of working. Similar to innovation, our progress on tech augmented ways of working will be evident in the performance programs I share. Under the functional scalability stream, we have 3 programs supported by projects with an ambition to deliver up to 80 basis points of cumulative margin improvement from 2023 to '28. The 3 programs under this stream are shared service centers, strategic procurement and overhead management. From '24 to '26, the focus was on securing gains through optimization. And from now to '28, the focus is on realizing the benefits from our structural investments. Moving now to the results achieved. This slide shows our margin gains to date from the programs. We have achieved approximately 110 basis points, as Hinda has already shared, towards the ambition we shared in 2024 of the 180 basis points by 2028. Of this 110 basis points, operational leverage has contributed 50 basis points. And the remaining 60 basis points have come from functional scalability. Just past the midpoint of the LEAP I 28 strategy, I can tell you delivery is on track. From '24 to '26, our focus was on capturing early optimization wins while laying the foundation for more structural gains in the second half of LEAP I 28. By 2028, we will deliver the remaining 70 basis points as our optimization wins compound and our structural investments bear fruit. In the next 3 slides, I will step into the 3 programs that contribute to our current progress and to our future 180 basis points ambition. I will start with functional scalability. Functional scalability has been the largest contributor to the gains we have made so far. The program has delivered 60 basis points from the 110 achieved to date. The program combines 3 connected actions. First, we have maintained disciplined overhead management by ensuring functional costs grow slowly -- slower than revenue. Second, we launched strategic procurement projects such as supplier base reduction and spend consolidation with suppliers. And third, we created and deployed the shared service center model, starting with IT across 4 global hubs. In IT, we have achieved 32% of our workforce being centralized, and we have achieved 30% reduction in our payroll costs. The structural work ahead is to complete the IT deployment and expand the finance and HR functions into the shared service center model. The model covers 43% of the workforce in the shared service workforce. I will now turn to the 2 programs under operational leverage, starting with performance management. The first program under operational leverage is performance management. This program contributes approximately 30% of the total cumulative 50 basis points delivered to date for operational leverage. The objective was to create a consistent data-driven operations discipline across the organization where we had inconsistent local solutions. We created standardized KPIs, a common global dashboard and a performance methodology that has helped managers turn data into decisions. The scale is significant, 63% of the global production workforce is in the scope of the performance program. As an example, in one midsized business unit using the methodology, we achieved a 24% improvement in net production per FTE, per field technical engineer. This is a good example of the methodology working as designed, and it continues to replicate across the globe. Scaling the program across all product lines and management layers is a key contributor to our total 180 basis point improvement ambition by 2028. The second program under operational leverage is process improvement, which contributes the remaining 70% of the total cumulative 50 basis points being achieved today for operational leverage. Process improvement is about building and deploying digital operation systems for our 5 key workflows across a large portion of the company. The program covers 58% of our global workforce. These digital operation systems standardize how we work and how we reduce our costs to serve. An added benefit is that the programs create the digital landscape for an AI integration. Approximately 60% of these digital operation systems are deployed to date, and the system will be fully deployed by 2027. As deployment scales and the system enhancements compound, process improvement will continue to drive the contributions from operational leverage. I will make this program tangible by sharing an example of one digital operation system for the certification solutions workflow. So SmartCert is an end-to-end digital workflow management system developed for our certification solutions product line. It is one of the clearest illustrations of what our digital operations systems deliver at scale. The system integrates the full certification solutions workflow from sales to invoicing and everything in between in a single integrated digital operations system. For the clients, the system also provides live interaction and project visibility. For our teams, it reduces manual tasks by 43%, reduces by half the number of systems they previously needed to access across the workflow, and it increases our capacity by about 15%. This is exactly how process improvement and other digital operations systems create value by standardizing workflows, simplifying the employee experience, improving client deployment into measurable operational gains. At full deployment by Q2 2027, the project aims at delivering 150 basis points organic margin improvement compared to 2023 by 2028. The harmonized workflows and digital operations systems like SmartCert also enable modernizing our way of working through -- further through AI enablement. So our digital operation systems like SmartCert provide the foundation for AI, which is already moving from experimentation to providing operational value. We see this happening at 3 levels across the company. First, more than 22,000 AI assistants have been created by skilled employees to improve their day-to-day work. Second, more than 300 employee-created assistants have been validated and published for reuse across the organization. And third, more than 140 solutions -- AI solutions are integrated directly into our workflows. At this level, AI becomes part of how we work, creating operational leverage at scale. Two examples bring this to life. In Marine & Offshore, an AI-powered knowledge management system covering approximately 15,000 pages of complex rules and standards strengthens the expertise, sharing them across 2,000 experts. In Consumer Product Services, an AI integrated into our digital lab operations system automates the ingestion of approximately 2 million client requests forms and has saved an estimated 150,000 hours. You will see more examples of AI-enabled projects later in the digital and AI session and in the tech corners and deep dives. This reinforces the performance story because AI is becoming the next layer of structural value creation on top of our digital operations system. And as Hinda has already mentioned, our next frontier is to reimagine and reengineer the end-to-end workflows with AI. At the midpoint of our LEAP 28 strategy, we have delivered tangible progress at scale, including 110 basis points of margin improvement. These gains have come primarily from optimization efforts with functional scalability making the largest contribution. At the same time, our structural projects are in process and are modernizing how we work as we move to digital workflows and begin to integrate AI-enabled operations. We have a clear path towards 180 basis point ambition by the end of 2028. The next phase is about scaling our programs, accelerating our adoption and extracting full value from the systems capabilities that we have built. Of course, performance excellence cannot be achieved through process, systems and technology alone. This is why the evolved people model is so essential. Thank you for your time. Maria, over to you.

Thank you, Noor. After so many numbers, I'm going to give you a break, and we're going to talk about people. I'm Maria Lorente Fraguas, Chief People Officer at Bureau Veritas. I joined the company about 2 years ago with 25 years' experience in the energy sector. I have an engineering background, and I spent half of my career in product development and operations roles internationally before transitioning to HR. And this combination of operational experience and people leadership allows our people function to act as a true connector between strategy and execution. And this is particularly relevant for us at Bureau Veritas because our success relies on our people and their expertise. Noor has explained how we are executing and amplifying our performance programs. But what does it take to sustain this performance? It takes having the right organization, strong leadership and the right capabilities. And this is the role of our Evolved people model. It is the pillar that connects portfolio strategy and performance execution, ensuring we build and deploy the capabilities to win in our chosen markets and to sustain also performance at scale. In 2024, we made 2 commitments to develop new economy skills in areas like sustainability, energy transition and digital and to embed augmented ways of working across the organization, helping our people work more effectively and focus their expertise where it creates the greatest value. Two years on, those priorities are as relevant as ever, and we are on track with the programs well established to deliver them. So let me share with you the progress we have made in our programs, placing the experts at the center of our people strategy and building on what makes us unique, the scale and depth of our expertise. As Hinda mentioned earlier, we are a global multi-specialist company. More than 64,000 technical experts work across 10 product lines with a wide range of expertise distributed across more than 90 different technical domains. And these figures illustrate not only our scale, but also the breadth and the depth of our expertise, the expertise that we can bring to our customers because our experts understand the industries in which our customers operate, the regulatory environment they must navigate and also the technical risks they need to manage. So our opportunity is very clear to make this expertise more visible, more connected and easier to deploy across the organization. And this becomes even more important as the world of work changes around us. In 2022 -- 2024, we also highlighted 2 major shifts: the growing scarcity of critical skills, and the emergence of technology augmented ways of working, and both shifts have now intensified. Critical skills remain scarce and are evolving rapidly in areas like sustainability, cyber and AI. And these are not just peripheral topics to us. They're increasingly central to the services our customers require from us. So recruitment alone is not enough. We must develop these capabilities internally and at scale. At the same time, AI is changing how work is done. It's changing the nature of the work itself, but also the technology our people use and it's also changing the capabilities that our experts, managers and leaders need to have. And for us, as we mentioned before, there is an additional dimension. We need the AI capabilities to transform how we work internally for sure. But we also are building the expertise to assess these technologies and give customers confidence in their use. So we have a dual responsibility to use AI effectively and to build trust in its use. And because all these changes are happening at an unprecedented pace, our ability to learn and adapt will increasingly determine our ability to lead, which is why the priorities we set in 2024 matter even more today. So over the past 2 years, our focus has been on putting the foundations in place. We started with the operating model because the organization and mindset had to evolve before everything else could. And we began to build the capabilities we needed to support our growth priorities, including programs such as the sustainability graduate program. And together, these elements provide the organizational, leadership and capability foundations of our evolve people model. I will cover each of them, starting with the operating model. And we started with this because to become a multi-specialist company, we need an organization that makes it easier to develop, connect and deploy expertise. As Maria explained earlier, our portfolio strategy has 2 priorities: focus our portfolio on the most attractive markets and extending the reach of our specialist product lines across customers and also geographies. And our operating model is enabling both because it combines a product line-centric organization with a shared leadership framework. Put it simply, the organization provides us the structure and the leadership framework, the behaviors that make it work. We have introduced a common structure across the company built about a simple principle, geographies own the market, product lines own the expertise. And this makes it easier to navigate the organization, connect expertise across the group and also deploy it where it creates the most value. We are already seeing the benefits of this new operating model. On the business side, stronger collaboration within but also across product lines has contributed to a significant increase in our pipeline. On the people side, our teams are benefiting from more opportunities to learn, grow and take on new challenges across the company. And we see this reflected in our employee Mobility Index, for example, which has increased from 17% to 23% since 2023 and also our retention rate, which has improved by 100 basis points. Our focus now is to deepen adoption and convert that momentum into stronger growth outcomes. But of course, a change in operating model does not happen just by changing the organization. It works in practice only if the mindset and the behaviors evolve with it. And this is what our new leadership framework is all about. But instead of describing the framework in detail, let's hear it directly from our employees. [Presentation]

No matter how many times I watch this video, it still brings emotions because what you saw in the video is actually our employees working in our facilities, in the customer premises and probably expressing better than we could ever express the leadership behaviors that we need to embody. So this brings me to the next priority of our people strategy, building the new economy skills that our portfolio requires. We have always been highly effective at recruiting technical talent. But in a market where critical capabilities are scarce and evolving rapidly, external recruiting alone is not enough, as I mentioned before. And so we decided to take a very different approach. We're developing those skills internally at scale. And the sustainability program is our first proof point of this model. It is a 15-month structured program designed to attract, fast develop and retain early career professionals. Participants benefit from an intensive learning experience and retention stands at 97%, but the value of the program goes way beyond the individual participants because it brings together managers, mentors and technical experts from different product lines and geographies. It mobilizes our entire network of sustainability, strengthening that capability, and it helps us connect expertise across the organization. We now have more than 3,000 sustainability experts supporting our growth in transition services, and the program gives us a repeatable blueprint that we can progressively apply to other priority areas like cyber, mission-critical or renewables. So with the organizational leadership and capability foundations of our Evolved people model now in place, the Amplify phase is about getting more value from them. And this is where our Expert-First approach comes central. At Bureau Veritas, expertise is our differentiator. So rather than starting with technology and asking where we can deploy it, we start with our experts and ask different sets of questions. What is their judgment creating the most value? What work can be simplified, automated or supported? And how can we give them more time to focus on customers, complex decisions and technical expertise? So our goal is simple, to amplify the impact of our experts. And this is why one of our key priorities in the Amplify phase is building a tech-augmented workforce. This builds directly on the work Noor described earlier. We are modernizing workflows, simplifying processes and embedding digital tools across the company. The objective is to remove friction around our experts, improve how work flows across the company and give people better tools to do their jobs. AI is definitely the next step in that journey. It will allow us to go beyond digitalization and begin augmenting expertise itself. And while AI is already part of how we work at Bureau Veritas, we're just at the beginning of what's possible. The opportunity is not just to give people access to AI tools, it's to build genuine capability. The confidence, the judgment and the skills to use AI in ways that strengthen what experts can do, but also in a way that our customers can trust. So to do that at scale, we will be developing different programs for different populations. We have designed 4 targeted programs to address this. AI Spark builds responsible AI adoption across the company, moving people from AI-aware to AI-enabled and ultimately, AI-ambassadorship. This is a program for the entire workforce, so nobody is left behind. Build the builders equips process owners and transformation leaders to redesign workforce with AI. And AI for Developers supports our development and product teams in applying AI to their work. Last but not least, leading a tech-augmented organization prepares senior leaders to prioritize investments and also lead adoption at scale. And the early results from AI Spark are really significant. We have 3,000 people that have completed the training. We have 10,000 weekly active AI users. And as Noor mentioned, we have also more than 22,000 AI assistants that have been created in-house. Our belief is that AI will amplify expertise that differentiates us, not replace it. And let me conclude by bringing everything together. We are a global multi-specialist company across 10 product lines and more than 90 technical domains, and our differentiator comes from the scale, the depth and our customer proximity of our expertise. Our people model is designed to maximize the impact of that expertise, and we are doing it through 3 mutually reinforcing enablers: First, a new operating model and leadership framework designed to scale expertise. It gives us the structure, the speed, the behaviors that we need to connect the expertise across the group and deploy it where it creates the greatest value. Second, building the new economy skills required for our growth priorities. We have achieved this through our sustainability graduate program, and now we have the blueprint to replicate it in other growth priorities. Third, tech-augmented workforce. We have put in place the program to modernize how work gets done, simplify workforce and equipping our people with digital and AI capabilities that amplify what our experts can achieve. And this is what we mean by an expert-first company, a company that can scale expertise across multi-specialist organization, a company that can build critical capabilities required for future growth and a company continuously increasing the value created by experts. Thank you. I will leave the floor to Philipp. He will cover Digital and AI.

Thank you, Maria. So good morning. I'm Philipp Karmires. I'm the Chief Digital and Innovation Officer for Bureau Veritas. I'm responsible for group IT technology, digital AI and innovation. I joined the company at the start of 2024. My passion is enabling productivity, growth and innovation through technology. Over the next few minutes, I will show you how we have built the group's digital foundation, how AI is already improving our operations and how we intend to convert these capabilities into productivity, operating leverage and new sources of value through 2028. Our digital innovation strategy is built around 3 connected layers, the digital backbone. I first announced it in 2024 as part of our LEAP 28 strategy. The goal of the strategy is to establish centralized technology standards, supporting our large global setup, as highlighted by Maria, with our 82,000 employees operating in over 100 countries. The first is enterprise foundation, the common infrastructure, governance, data capabilities and technology partnerships that support the entire group. The second is operation systems, the platforms embedded in the daily delivery of our services to clients. The third is innovation platform, where we integrate AI, create differentiated client experiences and build new service capabilities. The layers reinforce one another. The foundations enable our operation systems. Those systems generate the workflows and data on which innovation can scale. Together, they support 3 outcomes: greater productivity and scalability, technology augmented services that strengthen our competitive position and new revenue opportunities. From 2024 to 2026, our priority has been execution, establishing the technology and operating foundations required to scale digital and AI across Bureau Veritas. At the enterprise foundation level, we built a multi-cloud architecture, group-wide governance and shared technology capabilities deployed consistently across the business. At the operations systems level, we are building the platforms that run our core service lines. SmartCert is the clearest example, which Noor highlighted before, and I will cover with more technical details. At the innovation platform level, we built the AI integration framework that connects our operation systems to AI services in a controlled, governed and scalable way. This execution has moved digital and AI from fragmented initiatives to an integrated group capability. We are now ready to amplify its business and financial impact. The execution phase has delivered a technology foundation operating at a significant scale. Through technology standardization as well as shared services for IT, as highlighted by Noor, we moved from decentralized environment with applications and systems redundancy at a product -- to a product line-centric organization. In addition, we made sure we have AI infrastructure capable to interface with our systems at any level. Today, 95% of our applications are cloud-based, including our core enterprise platforms and operation systems. This gives us the flexibility to support growth, portfolio evolution and acquisitions without rebuilding infrastructure each time. We are doing this through working with leading technology platforms and innovation partners in a multi-cloud architecture. Across our systems, Bureau Veritas holds 8.8 petabytes of proprietary data accumulated through decades of inspection, certification, testing and assurance work. The opportunity is to progressively govern, structure and activate the most valuable elements of this data for operational improvements and new services. The foundations we have built and as well as the operational systems are important because AI is creating opportunities on 2 fronts for Bureau Veritas, as outlined by Hinda this morning. To recap briefly, on one side, AI enhances how we deliver trust. On the other side, AI itself creates new trust requirements. Let me explain the structure how we have -- the structure that we have introduced as part of our execution phase. We govern AI across the group through one framework with 4 connected components: deploy, modernize, activate and assure. Deploy and modernize are our immediate priorities. Deploy is about covering the large group and number of employees, as highlighted by Maria. It means providing the technology, controls and skills that make AI usable at group scale. Modernize means integrating AI into operation systems and transforming how core work is performed, which aims through our performance programs at 60% of our operations, as you heard from Noor before. Activate converts these capabilities into differentiated services and new client value. Assure extends our trusted third-party role to AI-related risk, governance and resilience. To learn more about Activate, please visit the Deep Dive for modernization inspection later this afternoon with Ginger and Bertrand. For assure, please visit the Assure New Trust Domains deep dive with Vincent and Pierre. AI is deployed across Bureau Veritas today at group scale with early operational impact. Deploy is about making it available for our applications and our people. For applications, we do this through providing internally an AI API, a programmable interface for internal applications and projects to integrate with various large language models. There are currently over 140 integrations through that option. For people, more than 10,000 colleagues actively use AI personal assistants on a weekly basis, including our own platform, that serves multiple large language models as well as through Microsoft Copilot. The compute capabilities for deploying AI and large language models is at enterprise scale. Our AI platforms process more than 650 million tokens every day. AI has moved beyond experimentation and is now becoming embedded in how Bureau Veritas operates. This deployment is translating into modernization impact in specific use cases to enable operational leverage as previously shown by Noor. We are now able to activate the data of 83,000 audits through our systems integration. Equipping our salespeople with AI-assisted copilots show initial savings between 2.5 and 4 hours per week. And in certification review, reported review time has been reduced from 90 minutes to 45 minutes. These examples demonstrate the progression from broad AI deployment to measurable operational impact. From 2027 to 2028, we move into the amplifying phase. The 3-layer strategy does not change. What changes is the pace, scale and business impact of what we have built. At the enterprise foundational level, the focus shifts to sustaining, optimizing and industrializing the platform at lower technology costs. At the operation systems level, we amplify AI integration in existing operational systems. Embedding AI is a key element in new systems development. At the innovation platform level, we create new sources of value through AI-native process transformation with inspection delivery as a primary focus. Amplifying means visible financial results, improved operational leverage, scalable delivery and differentiated service capabilities. Let's look at 3 examples for the upcoming Amplify phase for 2027 and 2028, which we are currently developing. The first is about integrating AI into existing systems. The second is about embedding AI as part of systems development. And third, how inspectors natively engage with AI. Let's move to the first one. Certification is a EUR 550 million yearly revenue base for Bureau Veritas. SmartCert is our new end-to-end digital operations system for certification delivery. It covers the full workflow, sales, auditor portal, back office and client portal, all in a single integrated platform. It is a key contributor to process improvement, as highlighted by Noor. It's already operating at meaningful scale. Approximately 60% of Certification Solutions revenue is in scope by end of 2026. More than 1,000 users are actively on the system today. More than 4,000 certificates are issued through it every month. More than 40,000 production dates are scheduled through it every month. Integrating AI now into this system will support our evolution. Please visit our deep dive in the afternoon to learn more. Inspection for industrial and energy assets represents a EUR 900 million revenue base for Bureau Veritas. Modernizing how we deliver it is a centerpiece of the amplification phase. The workflow spans 5 stages: configure, prepare, inspect, report and capitalize. We are digitizing this end-to-end workflow with tools tailored to have AI embedded. By 2028, our ambition is to have more than 80% of inspection activity digitized and AI-enabled. More efficient and consistent execution improves speed, safety and client experience. At the same time, digital workflows create structured, validated data assets that can support recurring data-driven services beyond the individual inspection. And one last example for an AI native enabled approach. Today, more than 3,600 inspectors globally spend a significant portion of their time documenting findings and producing reports in the field. With our AI-powered voice assistant, we are fundamentally simplifying that process. Instead of typing notes after an inspection, inspectors capture observations naturally through voice during the inspection itself. The experience is seamless and unstructured, allowing them to focus on the asset site or customer rather than administration. The AI then transcribes the conversation, extracts key findings and automatically matches information against our internal data structures and reporting standards. This transforms how raw voice input -- this transforms raw voice input into structured inspection content. Importantly, the inspector remains fully in control. The final report is reviewed and validated through a human-in-the-loop process before being issued to the customer. The benefits are tangible as part of our overall performance programs, reducing reporting effort by 10% through accelerating report turnaround times while improving consistency and data quality. As we look beyond LEAP 28, I would leave you with 3 -- I would like to leave you with 3 key messages: First, we now have the digital backbone in place. We have modernized our platform, strengthened cybersecurity, moved to the cloud, built our AI and data foundations and standardized technology globally. That foundation gives us scalability and productivity benefits today. Second, we are now entering the next phase, accelerating AI impact. The focus shifts to embedding AI into our core operations, whether it's inspectors in the field, laboratory workflow certification processes or knowledge workers, AI becomes a force multiplier. The AI enhances our expertise, improves productivity and increases service quality. And third, the real opportunity lies beyond efficiency. As AI strengthens and structures our data assets, it opens the door to new technology augmented services, stronger customer outcomes and entirely new sources of revenue growth. The combination of our global expertise, trusted position in assurance and a fully developed digital platform creates a unique advantage. We are combining it to the Bureau Veritas operating model in creating a platform for sustainable value creation well beyond 2028. In short, the foundation is built, AI impact is accelerating, and we see a clear path to continued growth, productivity and new revenue opportunities in the years ahead. Thanks for your attention, and I'll hand over to Francois Chabas to update us on finance. Thank you.

Good morning, everyone. I'm Francois Chabas, Bureau Veritas Chief Financial Officer. I'm very happy to stand in front of you to report on the progresses we've made since the last Capital Market Day. Before getting into the details, one figure already, which is good. We are 10% more in terms of attendance compared to 2024, which shows some interest into the company. And as you have seen, LEAP 28 is an ambitious program of transformation. It's been mentioned many times, both on the portfolio side, obviously, improving performance. Maria has been very clear on changing the -- transforming the people model. But this plan is not only about words, it's framed obviously by a set of clear and measurable financial commitments and a promise of value creation. So we are now 2.5 years mid plan, as Hinda put it. And we will see in the next few minutes, 2 things. One has been delivered and how it has been delivered over 2024 and 2026. So a phase we call execute. And thereafter, I'll spend some time to cover the financial perspective over the next coming 2 years, a phase that we're going to call amplify. So for now, who is running the script, you can pause for a second because during the Q&A session, something came to my mind and that with others during the coffee break phase, a thought I wanted to share with you. And more than a thought it's actually a phrase that is a phrase from a great contributor to the theory of economy that you all know, a man called Smith -- well, not Adam Smith, but Hannibal Smith from the A-Team for the French audience. A man who used to say, I love it when plans go as planned, [Foreign Language] and beyond the joke, this phrase illustrates one of the -- I hope one of the talent that the Bureau Veritas team you've seen here led by Hinda has, which is a passion for execution. We all love plans, but we value execution. And execution is about adapting to some conditions and the discussion with the analysts were very enlightening. The boxes, I could name a trade war starting in the middle of the plan. I could name energy crisis. I could name some divestments we've done sooner than expected. But what you've seen today and led by Hinda, the team has delivered. And that's what I hope I will be able to demonstrate to you in the coming page. So the next view is the one you've seen in Hinda's introduction. It's a busy page, right? But it reflects the seriousness and the holistic approach of our plan. Bureau Veritas is delivering on every single commitment we've made in March 2024. Organic growth still within our mid- to high single-digit ambition, 10% plus in '24, 6% plus in '25. Combined with the positive incrementals of our acquisition, we have delivered in line with our ambition in terms of total growth at constant currency. Margin-wise, there have been a lot of talk around the basis point improvement. We have demonstrated '24, '25 that improvements have been delivered year-on-year and each year. Hinda mentioned it in her introduction speech, Bureau Veritas is a highly cash-generative company. And here again, we've delivered. Our cash conversion has remained well above our 90% ambition, above 100% actually over the last 2 years with a leverage that stay within our commitment of 1 to 2x guidance. Returns, which are measured as EPS increase plus dividend yield, have reached double-digit levels across the period as promised. So again, execution is not a plan on the paper. It's already reflected in the results. And I will walk you through how we got there and why we are confident in the next phase. So usual bridge on the growth. This is what has been done from 2023 onwards with a projection by the end of 2026 in terms of CAGR. I think numbers are speaking for themselves. Organic growth, we will have delivered 7% to 8% CAGR over the period. But we are not simply growing. We are changing where the growth comes from. And moving to what Marios tell to you at the beginning, at the launch of the plan, we had identified the 3 famous boxes, portfolio categories, Expand Leadership, new strongholds, which were 2 categories where we're expecting the bulk of the growth to come from, and Optimized Value and Impact made of more mature businesses. So what has happened? I think we hit the target totally on new strongholds. New strongholds, which accounted for 10% of the revenue at the start of the plan, has generated 25% of the growth, fully 100% in line with expectations. And it's particularly remarkable coming from a business that used to represent 10% at the start. Now I need to use sentence to say, well, the main difference compared to our initial assumption lies in the respective contribution of our mature businesses versus our Expand Leadership businesses. And as you see here, with Expand Leadership, Building & Infrastructure contributed slightly less, especially at the beginning of the plan. We had B&I growing, some of you may remember, around 4%, 5% in the first 1.5 years. Good news, over the last quarter, the last 3 quarters, actually, strengthened by acquisition made in infrastructure, which are now delivering on organic after the usual 12-month period and the ramp-up of the businesses and Ciaran and Renato presented to you, the data center-driven assets, the contribution now is getting stronger and stronger. And over the last 3 quarters, we grew almost 9%. On the other front, the team eager to execute has phased and has been capable to leverage opportunities within oil and gas and within the second segment of this optimized value, which is marine, we've managed to recruit, staff, address clients in areas which were not initially the #1 in the priority of the company. So we managed here to get outperformance in the first years of the plan, obviously, supported by favorable investment cycle in those 2 markets. Looking at the inorganic revenue part of this bridge. M&A is a crucial element of our growth engine. We've gone through disciplined acquisition, which have added every year when you deduct acquisition from divestments, 2% to 3% in average CAGR over the duration of the plan. The portfolio rotation, of course, has partially offset the net increase. We have targeted disposals. Some of them have been done a bit earlier than anticipated. I think we shared a few times that the divestment of the food testing business was actually, in our mind, scheduled to happen in '26. We've divested it very early because we found the right way to monetize it with the right partner. So we've divested a bit ahead of plan so far. And then as most of the portfolio pruning is behind us, we now expect the net M&A contribution to be stronger in the coming years. And finally, FX remained a headwind, especially in '24 and '25. 2026 seems to be a different story. We see it to stabilize. So that overall, the 3 years will have a bit of a headwind effect on the top line. Now I think it's been mentioned a few times, we are moving towards a more interconnected product client-centric model. So it creates a new operating model that simplify growth algorithm, which will be music to the ears of the analysts. In simple terms, it brings more focus, better scalability, more simplicity. So as a consequence, Veritas has changed its way of reporting. From Q3 onwards, we will publish along those lines. Instead of 6 reporting lines, 21 subsegments, our performance is now articulated around 4 divisions and 10 product lines. No surprise here. This is what has been mentioned before. And I will spend now to share with you some elements which have not been disclosed yet to the market. I will see in the press release, but historical elements regarding the performance of those 4 divisions. What you see here is the same growth bridge, but along those 4 divisions. So what has not changed? Organic growth remained broad-based. It's not one single division that has brought everything. Industrial & Commodities is our largest contributor with roughly 45% of the organic growth with a CAGR, which is close to 9.5%. It's been driven by services related to investment in energy, typically renewable and gas. It's been driven as well by what we have developed throughout the very first beginning of our plan. Building & Infrastructure, as I mentioned, which is the second division, started somewhat a bit soft until H1 2025. We're now posting way stronger numbers. So some catch-up to come here in the next couple of years. Business Assurance, third division, had a smaller but yet steady contribution to the top line growth, mainly related to assessment for sustainability-driven purposes, which remains very strong. I think we are still hitting double-digit numbers on that segment, together, of course, with the cybersecurity assurance. Finally, the last division, which is Product Testing and Services contributes at a steady 6%, 6.5% compound annual growth rate over the period. We will move to the change of portfolio in a second. But as you can already see, the activities planned for exit, which is the tiny box, gray box on the right, which were here for the first 2 years of the plan have contributed only 4% to the total growth, which means they have not been growth accretive at all despite representing roughly 10% of the portfolio. Portfolio rotation on top of organic growth, which is the next pillar, has moved from ambition to execution. So since the announce of the plan, we have signed acquisition worth more than EUR400 million of annualized revenue and divested or exit or about to exit businesses worth EUR800-plus million. Together, EUR 1.2 billion, I think you've got this number now well in mind of rotation, which represents 21% of the total portfolio rotation from 2023 baseline, which is quite a measurable change. Our capital allocation policy aims at redeploying cash generated by these divestments towards activity with higher growth trajectory and higher return, obviously. We'll come back to that at the end. So now if we combine the simplification of the portfolio, the refocus on growth accretive segments, the M&A contribution Bureau Veritas has changed. I think that's something that has been clearly stated by many speakers here. First, the exposure to segments where we lead or where the growth trends are stronger has moved from 55%, which is the leadership and new stronghold to 85% as we speak. As a matter of illustration, the current Bureau Veritas portfolio post divestment scope already delivered a growth above historical average. Hinda mentioning in her introduction speech, roughly 60 basis points of additional growth compared to the previous portfolio. And you can measure by this that some of the segments are indeed a very accretive growth trajectory compared to the rest of the group. Now from a margin profile, same thing. This is the vision of Bureau Veritas by division now. Two divisions have margin ahead of the company average, Business Assurance, Product Testing and Services, while the other 2 are positioned at or slightly below group average. It's an average view for the last closed years, '24, '25. However, overall, our performance improvement program applied to each of the product line within this division, irrespective of the margin level. We believe that the high margin does not represent automatically a floor or max. We've seen a good example of this with the program centered around business assurance. As has been mentioned, we have an ambition to create margin incremental of roughly 150 basis points on a product line, which is historically already at 19%, 18%, 19% of margin. Moving now to the margin bridge. What have we done over the last -- so 2 years of actual and we took as a landing for 2026, the current consensus. Organically, Veritas has delivered 120 basis points of margin improvement in 3 years. Our performance program have contributed to the tune of 110 basis points. We have implemented several restructuring on top of it, especially in '24 and '25, focused at underperforming units, and those restructuring are now paying off as well. At the same time, and I think you got it from Philipp presentation, we have engaged into a series of digital-related projects to help streamline our processes and increase -- and to increase the billable time of our teams. And we have had, of course, along the way, our fair share of business disruption of all sorts, naming a few that you know the war here and there and various disruptions. So the key message here is -- we stick to the commitment we made at the start of the plan. We are generating margin incrementals. We reinvest them into improvement of system and processes in order to transform the way Bureau Veritas operates on the long run. These are no quick fix here. I think we try and make this journey by taking the upside when they are reinvesting them at the right time. So -- and each year through the plan, we are committed to deliver margin incrementals at constant currency. If we move now to the cash aspect of things and to capital allocation. Cash generation remains one of the core strengths of the company. Cash conversion stayed well above the 90% line every year, see it on the left. Our working capital requirements has further improved from 6.3% in 2021 down to 3.7% at the end of last year. So anticipating questions, I believe we will finish by reaching a floor here at some point, for those of you who have been here for long, we started at 10% plus. But unfortunately or fortunately, our credit management teams are still very eager to break this floor. So we are deploying currently a renewed and digitally enhanced system, collection system, cash collection system that will cover all our legal entities worldwide by the end of '27. We had one but a bit old fashioned. And we will see the results of this for sure in the course of the second half of '27 and early 2028. Leverage remained within the 1x to 2x guidance. So we are on the lower end, obviously. This gives us real flexibility to fund the next phase of M&A acceleration, as Marios and Hinda mentioned previously. Capital allocation. So our global framework, perhaps we were wrong from the start, is unchanged. So that's the same one you saw in March 2024. The M&A represents an important component. We have here on the page between M&A investment and divestment proceeds. And by the way, when it comes to proceed, the proceeds of divestment of our oil and petroleum activity is not yet there as it planned to arrive in the first quarter of 2026. CapEx remains targeted at lower end of the 2.5% range. Dividend ratio remained at 65% of adjusted net income, no change on that front. And within that framework, we have recently stepped up M&A, including the acquisition of LotusWorks. Now I think we've mentioned the necessity and the willingness to focus our effort in the next few years on M&A. And I would like to mention that discipline and acceleration are not in conflict. The framework -- this framework is yet built from the start to allow both to happen. A good way to illustrate this is to have a look at the evolution of our return on invested capital. At Bureau Veritas, we like clear slide and simple slide to read. This one has these features. Capital discipline is translating directly into clear and visible returns. The return on invested capital of the company has risen from roughly 16% in 2017 to almost 25% in 2025, consistently above our roughly 20% through the cycle. The beauty of this chart is that it demonstrates that the LEAP | 28 programs are directly generating higher returns. The company goes through many transformation, many changes. But all along the way, we make sure that on the execution part of it, we keep on delivering year-on-year and improving the return of the company. Now when we look a bit ahead and what's coming next in terms of Phase 2, we will be kind enough to move to the next page. For the record, there is no more script for the guy and difficulty to follow. Thank you for coming. If you don't move the page, you need to move the page. Good. Phase 2. So we now have the -- say, the full picture on what has happened. where are we going? And I'll spend a second on the next page, which reminds I think what Bureau Veritas has delivered so far, growth compounder, adjusted EPS growth and strong free cash flow. We don't intend to depart from that vision on the contrary. And we stick to the vision we set forth in 2024. That's why in terms of new commitments or updated commitment for the next 2 years of the plan -- next page. We -- 2 elements here, which are very important. We stick to the overall commitment of the plan. However, we precise 2 elements for the next 2 years, '27, '28. First, we upgrade the guidance in terms of total organic growth at constant currency to double digit. Two, we narrowed down the guidance in terms of leverage. We see that more M&A will come. So we'll be most probably in between the 1.5 to 2x leverage. And we have an ambition to reach by the end of 2030, sorry, EUR 1 billion revenue in AI-driven markets and services from a basis of EUR375 million by the end of 2026. Thank you for your attention, and I hand over to Laurent for the next phase. Thank you, Laurent.

So we have now seen the full picture, portfolio, performance, people, digital and AI. And of course, the financial ambition behind it all is a nice story on . Thank you, Francois. So it's time for the -- for our final Q&A of the morning. So all speakers are available. Please come to join me on the stage. And now that Francois has presented the numbers, everything is fair game. So we'll start with the room before moving again to our online audience.

So we're going to start, Rory, UBS, please.

It's Rory from UBS. My first question was around, I guess, the reinvestments you've been making. If about half of the benefits of the improvement plans are going back into the group, that means each year, you're spending about EUR 30 million already. Where is that going? Is that new greenfield investments? Is it new hiring plans going in to support future growth? But also how much of that is going back into maybe some of these transformation programs and spend that you're discussing?

Yes. Thank you for the question. I'll tag team with Francois on this. The -- a lot of the investments are going into this transformation. I think we were quite clear since this morning saying that the performance programs really need a complete reengineering of our operation systems, right? So that needs to be done. Our technology organization has actually done a lot of efficiency that they realized a lot of efficiency to absorb some of the costs, but we have to be clear. We have 10 product lines to modernize, so we need to invest on that. So a lot of it, if I have to break it down, but maybe Francois has a bit more color on that. It's really the modernization efforts of the operation systems. There are some efforts around new revenue streams. So there, you're really directly investing in the growth and development from -- for certain skills we needed to bring into the company. So that's the 3 kind of groups. Do you want to...

Really from a magnitude point of view, the bulk goes into the operation system. We are coming from a basis where all the foundation system were built now almost 10 years ago, but the operation system remains still a bit discrete. So in terms of magnitude, that's the bulk.

My second question is about the productivity gains that you outlined. What do you find your employees do with the freed up time? For example, if SmartCert freed up capacity by 15% for those engineers, do they have a backlog that they can go and work on more quickly? Do they do more time on innovation? Or do they just spend more time shipping to AI agents?

Yes. So it's a mix of things, right? So the capacity gives us -- the freed up workforce gives us capacity, so we can go after and get more work completed through the backlog that we have. We can spend the time on training as well and bringing the people, the expertise up. I mean those are the main things that, of course, if the work is actually not there, the backlog is not there, we can save some money.

I think the -- just to add to that, thank you, Noor. I think it's really important to understand that we are -- by freeing up time, we're actually rechallenging the workflows we have, and we're only starting, right? What we talked about so far is keeping essentially how we work. We're just making them a little faster, a little clearer, the system is more integrated, et cetera. And this is step one before we go journey-centric where we will completely redo it, meaning we won't have the same steps. We'll deliver the -- whatever the service for the customer, but how the inspector or the auditor is going to work is going to change. So what you -- the best way to think about this is a complete change of operating model for that particular service. So what work is done in the back office will be changed. So it's not only about the one person, it's about the number of people who participate in a particular operation. So in the SmartCert example, it's very interesting because what you're doing, normally, you do a lot of the scheduling, you have a bunch of people who sit in the background who actually do all those interactions. As you put a SmartCert system in place, they're actually freeing up their time because now the scheduling is becoming a lot more automated. And therefore, these people you can do 2 things. You can potentially redeploy them so they can do customer care and things like that or you might actually decide that you're not going to replace attrition, so you start reducing some workforce there. So it really goes beyond the one person, just to be clear.

François Digard from Kepler Cheuvreux. Maybe a follow-up. Is it fair following what you just said that the direction is that thanks to AI, it will allow you to grow organically at iso-FTE in the future? And connected to that, LEAP | 28 has been designed before AI. And nevertheless, the target on operating margin increment is the same. Is it because it is too early, step one, as you said, or because you are going to reinvest more than you previously thought or maybe that you suffer from or you fear some pressure on price?

Let me answer the second question, and then I'll pass to Maria to answer the first one. It's really important for us to integrate AI in a deep way. And therefore, it's a bit early today to fully account for the impact of AI in our workflows. So we didn't revise our ambition because we are only at the start of this. And you are absolutely right. When we started LEAP | 28, I think we looked -- we mentioned AI once in 2024, okay? So just -- and it's a fact. I think now that we are realizing the potential of the technology is changing our calculus on how we reengineer our workflow. So yes, it's not there because it's too early, but we're not shy to say that beyond the 2028, we see upside. And a lot of the work of the teams now in the performance team with Noor and Francois is really to look at that potential beyond that. As to the pricing, look, it's early in the journey. We haven't seen really customers talking about pricing yet. What I see and I see customers quite a bit, they're actually asking, can you adopt AI as fast as you can so you can help us with our performance and you can essentially move from being important, but a bit of a hindrance to critical and a performance enabler, which is a completely different conversation for someone in our space. So it's -- we're not there yet, but we are prepared that we need to protect the value, if that's what you were alluding to. Maria?

Yes, sure, sure. On the impact on AI on the workforce, I mean, we don't see AI as a headcount reduction story. We see it as, on one side, augmenting capabilities. On the other one, augmenting the impact of our experts. And then third, actually, it's true that we are hoping -- we are seeing already that we will be decoupling, let's say, the growth from the headcount requirements, right? So it will be decoupling. And we are right now looking at some modeling and assumptions. And it's definitely -- we think that about 75% of the requirements of headcount may be absorbed by AI, but not only AI, also the performance programs that Noor explained earlier.

Francois?

Just on the margin discussion, it's a delicate balance between the level of investment you commit and the output you can really naturally see in your P&L. So we do it a bit in a good old fashioned way by incrementals. We select a couple of, let's say, verticals where here we invest and we can measure rapidly the impact because it's been mentioned, you save time here and there. To give you an example, if you save time of an auditor for, say, 3 hours in a week, you will not be able to do more audits, right? Because those audits are calibrated to last at least 1 day. But if you free half an hour of a new time because now it doesn't need to redo the reporting CF and is done to something else, then he can start to prepare the next audit. So it's a fine mechanic. That's why we don't commit yet at the moment on incrementals. But I think what Maria mentioned is very true. Decoupling the headcount from the top line is one of the ambition we have. And that's for an inspection company like Bureau Veritas, very important. Two, let's not forget, we -- most of the time, we do not sell man hours or man days. We are selling an opinion. So the client pay for the opinion. It doesn't pay for the volume of work. And I think that's one of the critical factor Khurram, our Head of Sales is repeating all the time that do not sell days, guys, you are seeing opinions. And this is very true in terms of protecting ourselves against price erosion. But we will report later on, on margin incrementals when we have something solid. I think the tradition is to report on stuff when we have measurable elements, and we do so most probably in the coming years.

And perhaps to wrap this up, I think to the question from that side and yours is I think we are at the stage in this industry that the ones to invest to completely remake how we deliver services and essentially the trust that is -- that people need needs to change because the technology is absolutely transformative. And we have the scale to do that very, very quickly. It is a very fragmented market, a lot of small players, some medium ones and very few large ones, as you know. And I think we have the advantage of scale, and we need to leverage it now. So that's why the investment is critical, and it needs to be done now.

Next question is coming from Annelies.

Annelies from Morgan Stanley. Just going back to the AI-driven markets and new services. So that bridge from the EUR 370 million to the EUR 1 billion, how much of that do you think can be delivered organically given the growth in some of those end markets? And then how much through M&A? And as related to that on -- I assume a lot of that work is very margin accretive. So is that also part of your margin expansion?

Sorry, I didn't catch the last part.

I would assume a lot of the AI-related work is margin accretive. Is that right? And therefore, is that also part of your -- it feeds into the margin expansion?

Yes, I'll talk about that. Do you want to cover the market?

On the component of the growth, it's primarily based on the existing set of activities we have, what we call critical assets, both data center and semiconductor and the existing capabilities we have in cybersecurity assurance. There is a component of M&A, but I think the bulk of the plan is organic because I think we've completed that has been demonstrated with the combination of LotusWorks and our pre-existing data center commissioning activities that we have now a good, let's say, platform here to grow. So you could say there is -- there will be some M&A, but nowhere close to explaining the bulk of the growth, not at all.

Yes. Just -- and the best way to think about is the EUR 800 million on the AI-driven market is essentially mission-critical and we have a baseline and the growth rate. The EUR200 million is digital assurance and a lot of that is going to require some M&A. We have decent growth on the cyber business at double digit, but we need to really acquire some competencies in AI assurance. So that's really where we're looking. And generally, just to give an idea on the market, those are small outfits. So it will be things that we can -- we just need to move fast on them, but they are small bolt-ons in a way.

Okay. And my second question is also on the margins. You talked about continuing to roll out performance and people programs across all of your new 4 divisions. Where do you -- across those divisions, where do you think the biggest opportunity is to expand margins, again, thinking about your 3 buckets and how those feed into those 4 divisions?

Right. I think we talked about Business Assurance today. Business Assurance is where we're making commitments, not because we don't want to make commitments on the rest is where we have advanced programs in transformation. And the SmartCert example is the operation system. I think what you will see then we talked about inspection, which tends to be very large in our B&I business and very large in our Industrial and Commodities. You can think about it in that order. Business Assurance is probably the first one to see the impact. And then we should -- we already are improving actually margins in B&I. And then I expect that Industrial and Commodities comes after that because we're going really AI native in terms of how we do make that change. Do you want to be more specific...

Suhasini from Goldman Sachs. Just a couple for me, please. Just to follow on, on that margin question. Obviously, your margins at the group level are going to be a combination of mix along with the potential for incremental savings, leverage, et cetera. And looking at your mix in I&C, B&I, you do have margins which are at group average margins or below. And depending on the growth rates, that can change your group mix. So how do you plan to account for that? And maybe do you need to maybe allocate capital in some of these divisions to accelerate the margin expansion potential a bit further faster?

Thank you, Suhasini, for this simple question. Well, it's been the question of the group for ages, right? And before this presentation, we've made the exercise once again with I saying, what is the mix impact from 2023, where we started the plan to today. I put aside the divestment stuff. I'm just taking the businesses which were here in '23 and the businesses which are in '26. The answer is statistically 0. You will be surprised, but it's been like 0 margin impact from pure mix over the 4 years. Not changed the way you put it. Now taking to your question moving forward, you're right, the 2 largest business of the company have a margin that is, let's say, one of them is below the group average, the other one is just on the group average. I think the answer sits into, one, improving the margin of Business Assurance. That's a very clear commitment has been put here. So a smaller business but highly contributing. And two, engage into a transformation of the inspection model that is very big in B&I and quite strong in industry. So the 2 other buckets, the large divisions have product line which will be really impacted by this. So performance is meant as being the very first driver. I would not put M&A really as a component here because I mean, I would raise the probability of us buying a company that does for real, not adjusted with 3-year savings, but for real more than 16% margin EBIT. I haven't seen many in all fairness. So -- and we will more probably favor companies doing 10%, 12%, and with cost containment and cost measure that we own that can bring this company to 16%, 17% rapidly with full control. So I would not put M&A as a margin accretive component. It's more, I think, towards more scale, more growth, margin will be a different story. So short answer, performance programs centered around inspection and business assurance.

And my next question is just on incentive program KPIs. Have you made any adjustments in the last 2 years to account for the differences in the portfolio, the changing portfolio structure?

Absolutely. Absolutely. Do you want to address that? Go ahead, please.

So definitely, I think the evolution in the operating model comes from starting with the organization, the leadership behaviors and then incentives, right? And maybe I can explain this with an example. So we are talking a lot about collaboration, cross-selling, scaling and neutralizing solutions. So you put the organization in place, the multi specialist product line-centric organization. Then you need to also -- in the leadership behaviors, for example, we have one that is called challenge the status quo, seek opportunities. Another one is drive collaboration, and they have deep meaning on what they mean for us. And then on the incentive, we actually redefined the structure of the incentives not only to put, of course, cross-selling as a KPI, but also we now have all employees that have the annual bonus have a part that is on what they can control, their scope, they call it my contribution and a part of the incentive that is on the team, the level. So you're basically structurally not only with the organization, but then the leadership behaviors and the incentives closing the loop to make that work.

We're going to take a question from the web, and that's on people again. I'm going to read the question. It's a very long question. So thank you for asking that way. So to give a bit of context, so it says that the last CMD, there was a discussion about shortage of skilled labor, not just across the group, but across the industry. Based on the presentation today, it appears that you are addressing that through internal training. So the question is as follows: how much of these internal measures alleviated the shortage of skilled labor and how much of it still remains that cannot be addressed that way?

Yes. I think just before I pass to Maria there, I think we need to step back. In 2024, we had still, particularly in the North American market, massive shortage of every technical capability you can think of at the engineering level. We're not seeing the same thing. We see some shortage for certain electrical engineers, certain AI specialties and so on. So the market of skills is moving as well. In fact, the shortage today is in skilled labor rather where there is a massive buildup. So it's just a point to say that the skills market is moving quite dynamically at the moment, and that's why we need to be in phase or rather aligned with that. But I'll let you comment.

No, I would just say that we're really taking a holistic approach, meaning that we are developing the programs internally. As I mentioned, with the sustainability program, it's not only the people that are participating in the program, how we're activating the entire ecosystem. We're also doing some acquisitions to acquire certain capabilities that we need. So we look at it holistically. And in the end, I think that as you could see, our retention is increasing. And it's a bit of everything. It's also we are getting now our employees have better opportunities to develop -- but not only that, they're participating in projects that are challenging. So we are experts first, meaning the experts, what they want to do is they use their expertise to solve technical complex problems. So we are giving them the opportunity to really have impact in what they do. And I think it's the entire ecosystem that is allowing us also to develop but also to retain capabilities.

Can we move to the next question from the audience?

It's Karl Green from RBC. I've got 2 questions. Firstly, just in terms of digital assurance and the M&A strategy there, how do you go about mitigating for obsolescence risk? I mean, clearly, AI safety is a super hot topic at the moment. The Frontier labs can't agree on what the right move forward is. So how do you go about assessing that risk yourselves? And also, would you consider other models for delivering AI safety, partnering JVs, sell-through, et cetera? That's the first question.

Yes. Thanks, Karl, for the question. Very good question, actually. So yes, I would say yes to both. One is, yes, we need to partner sometimes. To give you an example, we did actually partner with AWS, where we worked on a specific AI assurance methodology for customers of theirs that needed reassurance that as they onboard on their cloud structure, they're actually -- they have a way to be assured and reassured that they are in compliance, particularly with the EU Act. That's one example. The other example is going to be using certain accreditation. So for those of you who will attend the trust in the new trust domain, we actually just received our ISO 42K accreditation. That's very specific, and we are ramping that up. I think Vincent will talk about that later on. And then the more difficult bit is really acquiring the technical skills so you can assess, to your point, the risks associated with new Agentic AI capabilities, right? And that's really where we will be scouting for either -- and it will be small. It's not going to be massive companies. It's just -- it doesn't exist. It's people who have started putting together some structures where they can do that. Alternatively, we will have to ourselves hire these skills, and that's why the previous question is quite fitting actually because those skills we need to onboard to devise how we will do this kind of assurance. So it's fair to say it's a very, very kind of fast-moving space. But we cannot not develop skills in AI assurance because everything is adopting AI and you cannot stay at the periphery of that, if that makes sense.

Very clear. And my second question for Francois perhaps. Just in terms of the north of 20% ROIC target by 2028, you're clearly doing better than that already. Are you ruling out it dipping below 20% in any given year between now and then? I mean, clearly, there's only 2 years left? Or is that just a kind of through-cycle number, just to clear and clarify what you would accept in terms of potential dilution from a platform acquisition?

Well, thanks. I think the 20% is a good common sense number for a company like Bureau Veritas, and we would not wish to go below. We are not guiding on a precise number, but I think the chart is rather clear. We want to remain high in terms of return on invested capital. That's the way we manage the company. So we could foresee in the next 2 years that we will stay above that level regardless of some acceleration on M&A that usually on the very short term, slightly reduce this number. But I think we show our confidence on being very much committed to remain above 20%.

And perhaps that's an opportunity to clarify something I had a few discussions. The M&A program, as Marios explained quite well, is we have a bolt-on high-speed fast lane. We're going to do bolt-ons to address the -- and that's really the bulk of what we do. And there, we have a lot more control. We have proximity. We generally have the relationships, and it's a size thing. So we have a capacity to have a good exchange with these sellers. And then you have the specific strategic expansion, which will be what we call midsized, but they're not very big. And in fact, you have seen it with LotusWorks, right? So we have the levers in our hands to manage how we deploy capital for M&A, which is why fully supportive, of course, of what Francois just said in terms of our capacity to protect the returns.

Okay. We still have time for one more question, and I know that Victoria was looking for it. So go ahead.

My question is back to certification. So have the rollout of the SmartCert platform actually disrupted growth at all in the short term whilst you've been rolling out the system as first half growth here was a little bit slower. And in terms of the investments you're making in certification, can you give a bit more granularity on where those investments are, whether it's AI, the production system or the specialist headcount, kind of how you're appropriating the spend there?

Noor, do you want to address the rollout?

Yes, I can say a few words about it. So it has not affected the operations across the board. The deployment of the systems are very structured. We have teams that are supporting it globally. We have trainings that are conducted for our employees. It's a very structured program from the development to commercializing the models and then to deploy them in the field. We feel -- we have a good as well feedback from the teams on the ground in terms of how they perceive the system and how they are operating. So in general, I would say there isn't. There's always, of course, opportunities to continue to improve, and we do that. So with every deployment, we look and see how we have performed, identify ways to get better and reestablish our deployment model and development model and off we go. So yes, I think it's going fairly well, but I wouldn't say it's affected the business in any way. If anything, it's starting to help and is starting to already yield some of the gains that we are trying -- we are -- have talked about.

Victoria, if you meet an inspector of ours or auditor and certification that complain literally initially. But after that, they're very happy with it. So it's just to give you the 2 views. Philipp, do you want to address the investment on SmartCert, roughly how we are managing that.

Yes. So the SmartCert component is like a component operation system, which we launched with LEAP | 28, so like pre-AI as it was highlighted. So right, we did the initial investments in building out this global platform, which is one platform for all countries, for all inspectors around certification. So that's part of the investment. And now we're basically upgrading the system through integrating AI. But the way we're integrating AI is we have this AI platform, which we built internally where we have various large language models that are fit for purpose and we integrate them. So the cost is rather the usage, optimizing the token usage against the functions that we are offering, think about comparing documents, think about processing information faster. So that's how the investments are currently split up, but the initial investment is around the platform and rolling that out globally. And the AI is the upgrade for it.

Okay. So we are at the end of the Q&A, perfectly on time. So thank you for the questions. And let's welcome again, Hinda for closing remarks.

Thank you for your attention and for saying for this few minutes. You have now listened to the team's presentation. You have heard us answer some questions there on the progress of the 3 pillars of the strategy and how we enter the next phase of LEAP | 28. I hope we are -- we have passed the message that we have a clearer portfolio, stronger operations -- operating platform and really increased capacity for us to deploy capital with discipline where it creates the most value. I will start a bit with LEAP | 28. So just briefly, if we can go to the next slide, please. Our vision is clear. And I want to pause on that because it's really critical that we explain customer intimacy is actually where you can get premium in this sector. So we will continue to put the vision at the center of what we do for customer excellence and sustainability and being that preferred partner. As the risks increase, the complexity increase, the technology increase, that partnership is very, very valuable. We're not -- increasingly, we're going to move in a less transactional role, and I'll come back to that on the AI. Now this is -- the LEAP | 28 is making Bureau Veritas more focused in where we compete, more scalable in how -- if you can move these in how we operate and more disciplined in how we allocate resources. We are preserving what differentiates us, though, because this is really important, the independence, the expertise, the global reach, the trust while we're strengthening the growth and the performance profile of the company. The model now will deliver high organic growth, will expand margins. And of course, we will continue to pursue the strong cash conversion we talked about. We will invest in performance program. In fact, I should say we must invest in performance program in AI and technology enablement. It's really critical for us. We will also accelerate the disciplined and value-accretive M&A, and I hope I have convinced you that we are doing that, again, in a very thoughtful way, and we're giving ourselves specific targets on returns so we make sure we stay aligned with what we achieve. As we enter the amplification phase, one accelerator is becoming critically important, and that's AI, both as a market opportunity, and we're committed to that and as an operating lever. So AI is strategic accelerator, both commercially and operationally. Commercially, it is increasing demand in attractive end markets we already serve. We talked about mission-critical, the data centers, the semiconductors, energy, the other critical infrastructure. Operationally, it is going to really unlock performance, unlock performance a lot faster than just digitalizing and connecting networks. And that's by connecting experts the data and the system so we can deliver the services in a way that becomes performance inducing for our customers. And then strategically, and back to many of your questions, it opens new assurance opportunities. And those are critical because today, a lot of our clients are looking for trusted partners to manage AI-related risk that they themselves are starting to understand, but they need help with governance, they need help with compliance, and that's exactly our profession. That's why AI for us supports both sides of the equation, the growth, the expansion of revenue and the strengthening of our performance. The ambition of EUR 1 billion by 2030 from a baseline of EUR 375 million is real. It has plans. We will execute and we will report back throughout the coming years. To conclude, we can go to the next one. The starting point where we are today at this very moment is stronger organic growth driven by higher growth markets, clearer product lines and the scale of our multi-specialist model. M&A is a disciplined accelerator. It will focus on capabilities, on leadership position and on attractive end markets. The performance-led execution program is laser-focused on productivity, on simplification and on AI integration so we can support margin expansion that we will prepare beyond this plan. And because the model is highly cash generative, it gives us the flexibility to reinvest and to fund the value-accretive acquisitions I talked about. But of course, we will maintain the balance sheet discipline, and we'll continue to deliver the return. That's why we are confirming the mid- to high-single-digit organic growth guidance, and we are targeting the double-digit total revenue growth in the next 2 years, '27, '28. I believe that Bureau Veritas is entering the next phase from a position of strength with the strategy, operating model and the financial flexibility so we can compound value over time. Thank you very much. I will now hand over to Laurent, who I think will have to go through the agenda for the rest of the day. And for those of you staying, please, we have really a very interesting program, and I'm looking forward to continue the conversation with you later on today. Thank you very much.