Johann Marty, Audrey Bahlinger, and Frédéric de Lavenne attended the AWS Toronto Summit 2025. Two intense days, over 120 conferences, thousands of participants, and one clear observation: innovation has never been more accessible… but everything begins with data.

From the very opening, Andrew Kent Warfield, VP at AWS, set the tone with a powerful keynote address: “Thanks to AWS technologies, there are no longer any limits to innovation. Turn your dreams into reality.” This powerful message resonated throughout the entire event, echoing in every session and conversation. The technologies are here, the tools exist and are more sophisticated than ever, but one fundamental element remains unchanged: without reliable, clean, and properly leveraged data, innovation remains nothing more than an unfulfilled promise waiting to be realized.
This year, generative AI and Agentic AI dominated every corner of the summit. The conversation has evolved far beyond models that simply answer questions or generate content. We’re now discussing intelligent, autonomous agents capable of self-orchestration, making complex decisions independently, and executing sophisticated multi-step actions without human intervention. This represents a major evolution marking a strategic turning point in the industry: we’re entering an era where AI doesn’t just react to prompts—it takes initiative, plans ahead, and acts with purpose.
AWS showcased several key innovations designed to accelerate this transformation and make it accessible to organizations of all sizes. Bedrock AgentCore now enables rapid deployment of ready-to-use AI agents that can automate complex business processes with minimal setup time. S3 Vectors, meanwhile, simplifies data indexing and makes information immediately exploitable by AI models, dramatically reducing the time from data ingestion to actionable insights. The objective is crystal clear: reduce technical barriers that have historically held organizations back, simplify Proof of Concepts to make experimentation more accessible, and enable rapid, efficient adoption across the enterprise.
But one idea emerges with striking clarity from all the presentations and discussions: AI is worthless without properly prepared data. Building a robust and scalable data lake, eliminating duplicates, correcting inconsistencies, intelligently indexing information, and establishing proper data lineage—these form the essential foundation for extracting real value from AI models. Even the most sophisticated AI, trained on the latest models and running on cutting-edge infrastructure, will never deliver good results if the data fed to it is incomplete, poorly structured, outdated, or lacks proper context.
Most importantly, the key message from the keynote is unambiguous: the technology is ready and waiting, but everything starts with strategy. Today, we have all the building blocks necessary to leverage Agentic AI and transform our organizations… but the real question that every leader must answer is: what do we actually want to build with these powerful tools? Before deploying intelligent agents across the enterprise, organizations must carefully consider their workloads, identify use cases that create genuine business value, and orchestrate the various components in the right sequence, guided by a clear and compelling vision. AWS provides the technologies, services, and foundational infrastructure—the complete toolkit for innovation. The key to success lies in knowing how to assemble these pieces intelligently, with proper planning and the right guidance from experienced partners who understand both the technology and your business context.
Success also depends critically on establishing solid governance frameworks from the outset. Defining quality standards that everyone understands and follows, ensuring security and traceability of data flows across all systems, harmonizing practices between technical teams and business units, and creating clear accountability structures—these are all indispensable steps for building a data infrastructure that’s truly exploitable at scale. The challenge isn’t purely technological; it’s fundamentally organizational and cultural.
The summit also highlighted the critical importance of modernizing infrastructure to fully exploit AI’s transformative potential. With solutions like AWS Transform, migrating from legacy mainframes, VMware environments, or .NET applications to the cloud becomes simpler, faster, and less risky than ever before. But beyond mere migration, the real opportunity lies in creating more agile environments capable of breaking down data silos that have accumulated over decades, improving overall system performance, and laying the groundwork for more ambitious AI use cases that can truly transform how businesses operate.

Everything converges toward the same fundamental idea: before thinking about deployment and scaling, you must prepare the foundation. Data must be the absolute priority—not as an afterthought, but as the cornerstone of your AI strategy. Creating a robust data lake with proper governance, cleaning and structuring information with consistent standards, defining clear and effective governance policies that evolve with your needs—these steps represent the true starting point of any successful AI journey. Only then can organizations confidently launch targeted Proof of Concepts, test high-value use cases with measurable outcomes, experiment with intelligent knowledge bases that unlock institutional wisdom, and begin exploring the transformative promises of Agentic AI.
What emerges from these two days of intense learning and networking is the unmistakable sense that we’re entering a new era of technological possibility. AI and Agentic AI are fundamentally redefining our relationship with innovation, challenging traditional business models, and opening doors we didn’t even know existed. But the true revolution won’t come from the tools themselves, no matter how sophisticated they become. It will come from organizations’ ability to transform their data into tangible value, to build a solid foundation upon which these technologies can reliably operate, and to create cultures that embrace continuous experimentation and learning.

Ready for the future ?