The Microsoft AI Tour brought together industry leaders, developers, and decision-makers to explore the rapidly evolving landscape of enterprise AI. The event, featuring a keynote from Microsoft CEO Satya Nadella, showcased the company’s comprehensive vision for AI integration across business sectors and technological domains.
This ambitious program spanned the entire spectrum of AI implementation, from practical applications in public administration and healthcare to cutting-edge technical sessions on GitHub Copilot and Azure OpenAI Assistants. With a significant portion of sessions delivered in French, the tour demonstrated Microsoft’s dedication to addressing specific regional needs while maintaining a global perspective on AI innovation.
Download Now: 50+ AI Reference Cases
[Updated January 2025]
At the heart of the tour was Satya Nadella’s keynote, which not only positioned AI as the fourth major platform shift in computing history but also outlined Microsoft’s three-pillar strategy for AI deployment through Copilot, the Copilot AI Stack, and Copilot Devices. His presentation, which we’ll explore in detail, illuminated how Microsoft is architecting the future of AI-powered computing while maintaining a steadfast commitment to responsible development and widespread accessibility.
As a preferred Microsoft Partner for Premium Consulting, Solutions & Managed Services across EMEA, Devoteam was naturally present at this event to showcase its expertise in implementing Microsoft’s AI solutions.
Dimitri Cabaud, Tribe leader AI/Data at Devoteam and Microsoft Enthusiast, held a session called “Improve NPS and save time with an AI-powered Call Center“.
Sessions and workshops
The tour explored Microsoft’s AI ecosystem, emphasising practical implementation across various sectors and the transformative power of AI tools like Microsoft 365 Copilot and Azure OpenAI. The sessions covered multiple aspects of AI integration, from NVIDIA’s collaboration with Azure for enterprise deployment to specific industry applications in nonprofit organizations, public administration, sustainability, and industrial transformation.
Technical sessions dived deep into development topics, including GitHub Copilot prompting, Azure OpenAI Assistants, cloud-native infrastructure for intelligent apps, and customization of Microsoft Copilot through extensions and plugins. Several sessions were delivered in French, addressing specific regional needs in public administration, financial sector innovation, retail transformation, and healthcare advancement.
A significant focus was placed on the practical aspects of AI implementation and data readiness. The sessions covered SAP ERP integration, data analytics for AI innovation, and the fundamentals of generative AI in Azure. Throughout the sessions, there’s a consistent theme of responsible AI implementation and practical applications that can drive organizational transformation, with particular attention to the Azure AI Studio Model Catalog, which provides access to various models from providers like OpenAI, Mistral, Meta, and Cohere. The content was structured to support technical and business audiences, offering guidance on everything from basic AI concepts to advanced development techniques and enterprise-scale deployment strategies.
AI in Cancer Research: Institut Curie Adopts Copilot
Institut Curie, a leading cancer research center, is integrating Microsoft Copilot to optimize its research and care processes. Facing the growing challenge of managing massive amounts of information, the institution has identified two main goals for its AI adoption.
The first aims to free researchers from time-consuming administrative tasks that divert them from their essential mission. By automating these processes, the Institute allows its teams to refocus on their true expertise: cancer research.
The second objective is facilitating and enhancing collaboration between scientific and medical teams. To illustrate this point, the Institute highlights the example of its team working on chemotherapy resistance in breast cancer. Researchers can now simultaneously query multiple databases and obtain relevant summaries in minutes, which previously took several hours.
From a technical standpoint, the solution relies on an architecture that allows Copilot to communicate with all of the organization’s data. The system can thus query different information sources, consolidate knowledge through artificial intelligence, and deliver clear, actionable results to users.
This technological advancement represents much more than process optimization: it allows Institut Curie to accelerate its research programs and experiments. By gaining this valuable time, teams can dedicate more effort to their fundamental mission.
Download Now: 50+ AI Reference Cases
[Updated January 2025]
Keynote : Microsoft’s Vision for the Next Computing Era
Satya Nadella reflected on witnessing four major platform shifts throughout his career at Microsoft. From his early days demonstrating Excel with ODBC connectors, through the eras of PC-server computing, mobile internet, and cloud computing, Nadella positioned AI as the fourth transformative wave in technology.
The Science Behind AI’s Exponential Growth
The driving force behind this new AI generation lies in what Nadella calls “scaling laws” – empirical observations similar to Moore’s Law, but potentially more profound. While Moore’s Law promised monthly performance doubling, AI’s scaling laws show performance doubles every six months. Nadella introduced a new equation for measuring competitiveness in the AI era: tokens per dollar per watt.
This exponential growth isn’t solely about computational power. It’s the convergence of three key elements:
- Advanced computing scale
- Sophisticated data techniques (both pre-training and post-training)
- Algorithmic innovations, particularly transformers (since 2019)
The Three Pillars of AI’s Impact
Nadella outlined three fundamental shifts that AI is enabling:
- Multi-modal Interface Revolution: For the first time in 70 years of computing, we’re seeing computers truly understand humans, rather than humans adapting to computers. The new interface combines speech, text, and video capabilities.
- Neural Reasoning: Moving beyond traditional relational databases, AI can now detect and predict patterns using neural algebra, making sense of our increasingly digital world.
- Contextual Intelligence: AI can now maintain context and additional memory, enabling rich agentic work through personal and organizational AI agents.
From Technology to Real-World Impact
Microsoft’s vision extends beyond technological advancement. Nadella emphasised that real-world impact will measure success: empowering small businesses, enhancing multinational competitiveness, improving public sector efficiency, and advancing education and health outcomes.
In line with Microsoft’s identity as a “tools platform partner company,” they’re approaching this vision through three key platforms: Copilot, Copilot AI Stack and Copilot devices. The foundation of this strategy is Copilot, which Nadella described as “the UI for AI” – an organizing layer designed to interface with an increasingly rich world of AI agents.
Download Now: 50+ AI Reference Cases
[Updated January 2025]
Autonomous Agents: The Next Evolution of Copilot
Building on the foundation of Copilot, Microsoft has unveiled its next significant advancement: autonomous agents through Copilot Studio. This announcement marks a crucial evolution in Microsoft’s AI strategy, demonstrated through a real-world implementation at McKinsey & Company.
Expanding the Copilot Ecosystem
Microsoft has launched over ten pre-built autonomous agents, covering diverse business functions from sales qualification to financial reconciliation. The key innovation lies in Copilot Studio’s ability to help organizations create their own fully autonomous agents that can either:
- Integrate directly with Copilot
- Function independently with standalone capabilities
- Escalate to human intervention when needed through structured workflows
McKinsey’s Client Onboarding Revolution
A demonstration featured McKinsey & Company’s implementation of an autonomous agent for client onboarding. This showcase illustrated how AI can transform complex business processes:
The agent automatically processes incoming client emails, analyzing content, matching against McKinsey’s industry taxonomy, and identifying appropriate partners – all tasks traditionally requiring human intervention. Most importantly, it drafts summary emails to engage the right partners with new clients.
The Power of Natural Language Programming
What makes this technology particularly revolutionary is its accessibility. Instead of complex programming languages, these agents are created using natural language instructions. The system comprises three key components:
- Triggers: Allowing 24/7 monitoring of specified events across IT systems
- Knowledge Base: Integration with various data sources including documents, SharePoint sites, and databases
- Actions: Leveraging over 1,500 connectors to automate processes across different systems
Intelligent Escalation
A crucial feature of these autonomous agents is their ability to recognize limitations and escalate to human oversight when needed. This was demonstrated through a real scenario where the agent encountered an outdated partner reference and appropriately escalated to a manager through Copilot’s interface, maintaining efficiency while ensuring accuracy.
This combination of autonomous operation with intelligent human escalation represents Microsoft’s vision for the future of work – where constellations of AI agents support individuals, teams, and entire organizational functions, streamlining processes while maintaining human oversight where needed.
Building the Complete AI Infrastructure: From Stack to Devices
The Copilot Stack
Microsoft unveiled its complete AI technology stack, making available the same infrastructure used to build Copilot. This platform spans multiple layers:
Infrastructure Foundation: Nadella emphasised Microsoft’s significant infrastructure investments, including the expansion in France with a €4 million investment in AI-accelerated compute. The company is innovating at every level, from silicon development (including their own chip, Maia) to partnerships with NVIDIA and AMD, marking what Nadella calls “the golden age of systems computing.”
Data Architecture: Recognizing that AI applications are inherently stateful, Microsoft has developed an AI-first data infrastructure. This includes:
- OLTP capabilities through Cosmos, SQL, and Postgres
- AI-first analytics infrastructure through Microsoft Fabric
- Robust data governance ensuring appropriate data usage and compliance
Azure AI Studio: The platform includes an AI app server offering:
- Broad model selection, including OpenAI and Mistral options
- Fine-tuning and RLHF capabilities
- Evaluation tools and guardrails for responsible AI deployment
GitHub Copilot’s Evolution
Nadella highlighted the remarkable adoption of GitHub Copilot, noting its transition from a skeptically-received tool to an indispensable development companion. In France alone, GitHub hosts 2.6 million developers, with 22% year-over-year growth. A demonstration showed GitHub Copilot with O1 preview optimizing its own tokenizer, showcasing AI’s ability to improve itself.
Copilot Devices: Bringing AI to the Edge
The third major platform announcement focused on Copilot devices, acknowledging that distributed computing remains crucial. The new Copilot PC initiative combines CPU, GPU, and NPU capabilities, developed in partnership with AMD, Qualcomm, and Intel to optimize both traditional PC performance and AI capabilities.
Trust Framework and Responsible AI
Nadella concluded by emphasizing Microsoft’s commitment to responsible AI development through:
- Clear principles for privacy, security, and AI safety
- Implementation of adversarial attack simulations
- Confidential computing for privacy protection
- Tests to address hallucination challenges
Commitment to France
The keynote concluded with a significant announcement for France: the introduction of the AI Skills Navigator program, building on Microsoft’s commitment to skill over 1 million people in France with AI capabilities. This initiative follows a documented 20% growth in AI skills since 2019, demonstrating Microsoft’s dedication to empowering French innovation and technological advancement.
Devoteam helps you lead the (gen)AI revolution
Partner with Devoteam to access experienced AI consultants and the best AI technologies for tailored solutions that maximise your return on investment. With over 1,000 certified AI Consultants and over 300 successful AI projects, we have the expertise to meet your unique needs.