So as you mentioned, I'm Laurie Barrington. I'm Schneider Electric's North America chief data officer.And it's a real pleasure to be here with you today to share a topic that's becoming increasingly central to how organizations operate, innovate, and deliver value in a rapidly evolving environment. During the next twenty minutes, I'm gonna focus on trusted data at scale and really the foundation for how AI how we deliver AI. I'd like to walk you through this journey by focusing on four key lessons that shaped our progress. First, the work that we did. So what are the foundations we put in place to scale and scale responsibly? The outcome, the impact our data first strategy unlocked. Third, the learnings, What did we discover along the way? And finally, what would we do different next time? Because trust evolves.

But before we start, I just want to take a second to introduce Schneider Electric in a few sentences. Schneider Electric is a global energy technology leader driving efficiency and sustainability by electrifying, automating, and digitizing homes, industries, and businesses. We are a €40,000,000,000 revenue company with over 160,000 employees and 1,000,000 partners in over 100 countries. We consistently ranked amongst the most sustainable companies in the world. A few years ago, well before the explosion of generative AI, we at Schneider Electric made a bold choice to accelerate AI solutions at scale. Our goal was and still is to enable our employees to work smarter, make our factories more productive, enhance the quality of our customer experience, and ultimately increase the value our customers receive from our solutions. But let's start with the core challenge we needed to solve, the never ending pilot mode. We have all seen promising AI experiments that never make it to production.

So how do we ensure that our AI ideation and experimentation actually reach industrialization? We started with scale in mind, not as a final step, but as the first requirement. But before that, we had some prerequisites. We needed to select use cases with tangible business value based on real customer needs, strong business outcomes, and strategic relevance. This already requires cross functional collaboration, business experts, customer experts, data experts, AI experts, all working together from the start. But even a strong use case cannot scale without mature data. So the second prerequisite is then data readiness. To build, train, and industrial AI, data must be available, reliable, consistent, and at scale. As we integrate AI across our offices, our products, and our services, one thing became obvious. Both AI capabilities and expertise and supporting data governance foundations were needed.

We cannot have AI without data. Data is the foundation and the trust layer for every single AI application. So how did we make this happen? By implementing a data and AI at scale requires far more than al algorithms. It requires foundations and disciplines and in intentional change. It requires organization and commitments. Our first priority was to build the organizational and governance backbone that would allow us to scale responsibly and securely. We structured ourselves around dual leadership. This model included AI and data working hand in hand. The AI office leads the AI strategy and the AI hub, which is a center of excellence of over 350 AI experts located in The US, in France, and India, and their mission is really focused on delivering AI value from ideation to industrialization. But as a chief data officer, my role is to ensure the data required for these AI use cases is trusted, secure, and accessible.

The organizational design directly feeds into our commitment. If data is used in decision making, it must be trusted from protection to creation to standardization and to the consumption in analytics and AI. Transparency being core at Cheddar Electric's principle, we formalized our commitments to build trusted data at scale. In our corporate trust charter, that is a statement of our responsibilities to all of our stakeholders, that can be found on our se.com website and where data appears not accidentally, but 74 times. And now you may ask, so how do we practically ensure trusted data? Sorry. This I'm going to have to apologize. I'm going to have to skip to a slide here, because it's not not building as in as intended. So I'm gonna bring it to the slide.

So what this is intended to talk about here is around our four data golden roles. Right? These are data strategy and governing principles. And concretely, what do they need? Golden rule zero is around data risk. The very first rule ensures that each project team has protocol to manage both event triggered risks, such as external complaints about privacy and retention, as well as data management practice risks. We protect our data by keeping it secure, private, and fully compliant with global laws and local regulations. This ensures we handle data safely and responsibly. Golden rule number one is around our governed authoritative sources. This rule reinforces the importance of identification and using only approved authoritative sources of data to secure its reliability and lineage, which in turn is essential for responsible AI.

This ensures everyone works with accurate information. Golden rule number two, governed group referential. This group refers to commonly used group referentials such as geography, market segment, customer classification. It aims to assure that the standard the the standardized definitions in common language for consistent and comparable analytics and enable unified understanding organization wide. And finally, golden rule three is really around our governed group platforms. This final rule makes sure we are use harmonized and reusable data objects from go from governed platforms. This reinforces that all data consumption can be traced back to a trusted source, enhancing the use of verified data for reporting and decision making. Next, let's apply these rules across the global organization where we deployed a strong network. So, this network is made up of business data offices, data domains, and data delivery teams. Our data offices have a mission of empowering business with reliable insights.

It's made up of functions, operations, and our businesses. Our data domains have a mission of really they're the guardians of the data for quality and security. And finally, our delivery teams really deliver data at reusable data objects. This network is essential because central teams alone cannot scale governance. By distributing ownership, we make data a shared responsibility and we raise maturity across the entire organization. We continuously measure and benchmark our data maturity. This allows us to track progress, identify gaps, and share best practices, and ultimately helps the entire company speak the same data governance language. I like to say data is everyone's business. Finally, we know that no governance model would succeed without a cultural shift. So we invest back to the correct slides here, so apologize for the incorrect build.

So we invest in education empowerment through our AI data, and AI school, learning paths for all of our employees, and then upskilling programs and expert certifications. So what were the outcomes? We'll start with our most critical and the biggest challenge for us as data offices. But as I shared previously, we kind of enabled this mindset of data quality by design in the company, embedding quality at the moment data is created through well defined business processes, clear accountability, and proper digitization. And when foundations are strong and trusted data is available, then scaling AI becomes not only possible, but natural. Faster time to value. Now that we have a robust approach to data cataloging, authoritative data sources, sensitive data classification, and controlled data sharing, including we have data sharing agreements that allow our AI teams to access data quickly and securely, and these therefore, they spend far less time searching for data or reconciling inconsistencies, they can then focus on what matters most, building and scaling.

Higher model performance using governed, high quality, consistent data has directly improved model accuracy, stability and resilience over time. Models retrain more smoothly, break less often, and integrate more easily into production systems. Thanks to this trusted data at scale, we now have industrialized more than 30 AI use cases every year, and the impact is visible across multiple activities. AI is no longer a pilot. It is an operating capability. We've operationalized AI use cases in the areas of manufacturing efficiency and predictive maintenance, supply chain optimization, customer support augmentation, energy management and flexibility solutions, and commercial excellence and pricing optimization. But what is some tangible impact that we've provide provided to our customers?

Let me take you through an example, about something that I think we can all relate to, and that's the home. Most of us already try to manage energy at home without thinking about it. Right? We turn off the lights when we leave a room. We unplug the charger when we're not using it. We avoid leaving devices on standby, or simply we just put on a sweater, to to instead of raising the thermostat. But today, our homes can do much of this thinking for us. As an example, our product, Wiser Home Energy Management with energy flexibility optimization, tackles this complexity behind the scenes. It helps households to draw less energy from the grid, to make the most of their solar production, to stay below contracted peak power limits, to lower their carbon footprint with minimal effort, and in some cases, even earn money by participating in demand response programs.

It's a simple illustration of what intelligent energy management looks like. Every life made more efficient through data driven, autonomous decisions without homeowners needing to consistently intervene. This is AI transforming sustainability, not in a lab, but in real homes. But now let's take a little bit of a step back and look at something a little bit more complex, a microgrid. If a home is a simple energy ecosystem, then a microgrid is the industrial scale cousin with multiple energy sources, loads, storage systems, weather dependency, and even market signals, and all interacting at once. In there too, Schneider has a solution called EcoStruxure Microgrid Advisor, With the microgrid control system brings all this intelligence together. With real time insights, it is possible to dynamically control on-site energy resources, automatically forecasting and optimizing how and when to consume, produce, and share energy.

What once required constant human oversight is now handled autonomously by a system that understands its data, anticipates what's coming, and adjust itself in real time. And thanks to innovation and effective AI technologies, EcoStruxure microgrid advisor has been recognized by WEF's global program, spotlighting AI application and delivers measurable impact across industries. And how do we make the decisions on which use cases to implement? As I mentioned earlier, we don't rely on intuition. We follow a clear, structured evaluation process that helps us understand the real potential of each use case for our customers. If trusted data is available and accessible for that use case is of paramount decision making. Of course, this journey from data to AI brought several important learnings. First, as I previously said, data is everyone's business. Without shared accountability across functions, neither data nor AI can scale. Second, data quality begins long before data appears.

It starts with business process design, clear definitions, roles, and governance. Third, access must be both secure and practical. AI teams need timely access to data, but within solid guardrails. Fourth, central governance alone does not scale. The network of business data officers was crucial to distribute this ownership. Fifth, AI demand actually improves data maturity because only use cases with high quality data are accepted. So teams naturally align on better data practices. And finally, education is a powerful accelerator. Without a shared understanding of data and AI, scaling would simply not be possible. What might we do different next time? If we started again or accelerated further, there are areas we would approach differently. We certainly would automate more, from data classification to lineage to quality checks to reduce overall manual efforts. We would engage business owners even earlier.

I talked about data starting right from the beginning with the process. So, when processes are being designed, we would embed quality by design into those processes. And finally, continuously evolving governance models to reflect all of these new regulations that are coming, these new technologies that are happening, and new expectations around trust. And at the heart of all of this is one belief that we have at Schneider Electric, and that's that partnership between data and AI must always be strong. So as a data officer, you can imagine that I love my data points. And so I'd love to leave you with three takeaways and three numbers. The number 74 is the number of times data is mentioned in the trust charter Schneider Electric's code of conduct helping transform our culture of data. 30 is the number average number of, AI use cases that we manage per year, really ensuring they're focused on strong business impact, but always with scale in mind.

And four, and this is foundational, and you can see kind of foundational to the other two things, is the number of data golden rules are key fundamentals in everything that we do to assure trust and scale in our data and AI implementations. Thank you.