Matt Harrison Clough/Ikon Images
At the World Economic Forum Annual Meeting in Davos, Switzerland, in January 2026, Schneider Electric CEO Olivier Blum accepted awards recognizing the company’s AI solutions as part of the WEF’s MINDS (Meaningful, Intelligent, Novel, Deployable Solutions) program — for the second time. The distinction highlighted two of the company’s AI-enabled applications: EcoStruxure Microgrid Advisor and SpaceLogic Touchscreen Room Controller — for delivering measurable impact in energy management. Schneider Electric is the only company to be recognized twice by the program.
“It is clear we have entered a new era where AI and energy are inseparable, and together they will reshape every business,” Blum declared at Davos. “AI requires compute, and compute requires energy. That is why the world needs greater energy intelligence.”
This interdependency between artificial intelligence and energy puts Schneider Electric — a global leader in energy management technology — at the center of one of business’s most critical challenges: powering the AI revolution while advancing sustainability goals. To meet both its customer needs and its internal process objectives, Schneider Electric has built an organizational model that deploys AI at scale, deliberately skipping the pilot phase that consumes resources without delivering business impact at so many companies. (More on this later.)
Philippe Rambach, Schneider Electric’s chief AI officer since 2021, is leading this effort. With nearly 100 AI use cases now running in production — split roughly evenly between customer-facing solutions and internal operations — Schneider Electric has demonstrated that AI can deliver value across every dimension of enterprise operations, including manufacturing floors, customer care centers, and complex energy-optimization systems. The company’s recognitions extend to manufacturing as well: In January 2026, the WEF’s Global Lighthouse Network awarded Schneider Electric’s Wuhan factory in China the Lighthouse designation — the company’s ninth — this time in a newly introduced category honoring talent development and people-centric workforce models.
Rambach described a strategy grounded in business value rather than technological experimentation. “We always start from the business and customer needs, pain points of employees, where AI can help,” he told us. Every initiative must demonstrate clear business value and plan for deployment at scale from its inception. Rambach and his senior management colleagues are also concerned about AI governance and ethics, but, as he noted in a 2025 report produced by MIT Sloan Management Review and Tata Consultancy Services, “Explainability matters — but in the boardroom, consequence matters more.”
Balancing Two AI Portfolios: Internal and Customer-Facing
Schneider Electric pursues AI opportunities across two distinct fronts, each with different strategic imperatives, approaches to measuring success, and timelines for realizing value.
Internal AI applications deliver more immediate financial returns, helping employees work faster and better while providing enhanced support for customers.
Customer-facing AI, meanwhile, represents a longer-term strategic play focused on capturing market share in emerging and evolving markets. “For customers, we want to be first to market and take strong positions, even in markets where AI-assisted energy management isn’t fully developed,” Rambach said. For instance, each country in which Schneider Electric operates shows different rates of renewable energy penetration and different challenges as large new electrical loads come onto power grids. This requires the company to adapt its AI solutions to diverse market conditions.
While much of the business world has rushed to embrace generative AI, Schneider Electric maintains a balanced portfolio of AI technologies. Analytical AI — traditional machine learning applied to structured data — still accounts for roughly 60% of the company’s overall AI work, particularly in customer solutions. “Analytical AI is very important and provides a lot of value,” Rambach emphasized. “We are not giving up on that.”
Generative AI represents about 40% of customer-facing applications and roughly 70% of internal, employee-focused tools. The technology excels at making systems easier to use and providing support capabilities, and at generating code, though Rambach stressed that significant human involvement in system development remains essential. Schneider Electric has also incorporated generative AI into its smart-grid solutions and is exploring the application of foundational transformer models to analyze internet-of-things and time-series data, and to create multitask models.
One of Schneider Electric’s most important applications of generative AI addresses a challenge common to large enterprises: making organizational knowledge accessible and usable. The company needed systems with robust security, clear information provenance, and the ability to cite sources. This required building vertical knowledge bases tailored to specific functions rather than deploying a one-size-fits-all solution.
The curation of unstructured data for these use cases proved instructive. “Asking people to clean their own data for data quality’s sake doesn’t work,” Rambach said. People are naturally resistant to what can feel like make-work. “But if you show them what you can do with it in an AI context, they are much more amenable,” he noted. When employees could see the direct impact of better data, they willingly performed curation work.
This insight reflects a broader principle at the company: Employees must be integrated into the AI development process. “People at the front lines are doing the work — they are at the core of Schneider’s approach to AI,” Rambach said. “We start from the business domain and bring in anybody else who is needed. Central experts don’t have the domain knowledge.”
Embedding AI, Not Building Stand-Alone Products
Schneider Electric deliberately avoids creating separate AI products for internal users or customers. Instead, the company embeds AI capabilities into existing systems and processes, such as energy management applications, field service tools, customer care platforms, and sales aids. A prime example is an AI-powered tool that it implemented for the company’s sales force, which must navigate an extremely complex product catalog. Rather than launching a stand-alone application, the company built AI recommendation capabilities into Sales Copilot. The company applies product management discipline to AI use cases, overseeing AI-powered processes and products from conception through deployment to eventual retirement.
This integration strategy extends to emerging capabilities like agentic AI, where Schneider Electric is already seeing practical value today despite the technology’s relative immaturity. The company has built an agentic system for processing requests for quotations that extracts key information, reformulates it, and summarizes it for salespeople. The system isn’t perfect, but it significantly improves sales productivity. “In many situations in companies, 80% to 90% accuracy is enough when there is human review,” Rambach noted. The key is educating users to review and improve the AI’s output rather than accepting it blindly. Schneider Electric is progressively moving toward more agentic process automation, shifting away from traditional robotic process automation while using AI as an adviser and recommender rather than a fully autonomous decision maker.
Schneider Electric takes employee understanding and behavior change seriously, implementing a tiered training approach that recognizes different needs across the organization. The company has made AI training mandatory for everyone but tailors the curriculum to four distinct groups. First, all employees, including those on production lines, receive foundational AI training. Second, management gets specialized training on leading AI initiatives and managing AI-enabled teams. Third, AI experts on Rambach’s team receive deep technical training. Finally, and most unusually, product managers, process owners, and IT owners receive training focused on how AI can enable transformation of their domains.
An AI Organization Designed for Scale — Without Pilots
Perhaps Schneider Electric’s most distinctive feature is its organizational model, which is explicitly designed to achieve impact across the organization quickly rather than generate pilots and experiments. “Our goal is not to have pilots and experiments: Use cases are deployed at scale,” Rambach emphasized.
This model rests on three components: a team of more than 350 people dedicated to AI; a comprehensive technical platform incorporating Microsoft Azure, Amazon Web Services, Databricks, large language model operations with retrieval-augmented generation, LangChain, and various APIs; and, perhaps most important, a structured process that guides initiatives from vision through ideation and incubation to deployment at scale.
At each gate in this process, the company confirms the business plan and business case. Success requires the merger of domain knowledge with AI expertise, which brings together product owners, IT professionals, data specialists, trainers, and marketing personnel. Some other organizations employ this stage-gate approach to AI initiatives, and we believe it is a useful way to increase the likelihood of a valuable outcome. It’s much more common, however, in new product development processes that don’t necessarily involve AI.
Measuring Value Without Waiting for Certainty
Measuring AI’s economic value presents challenges, particularly for customer-facing products, where the significance of advances is tricky to isolate. “It can be difficult in the customer product space to show value from tech improvements,” Rambach said. “What is the ROI from moving from desktops to laptops? What is the value of adding a communications protocol to a product?” The company does, however, track both usage rates and outcomes with AI-enabled products, such as energy savings achieved by customers.
For internal applications, Schneider Electric starts with a clear business value proposition and tracks two KPIs: an adoption target and a performance metric. The performance metric varies by use case; it might be accuracy, customer satisfaction scores, or a reduction in credit defaults. “One KPI in each,” Rambach said. “We follow whether it performs.”
Business stakeholders own the value proposition and help develop appropriate KPIs for their use cases. The company calculates total annual AI value and reports estimates to the board, projecting the technology’s impact over a four-year horizon, but it keeps these figures confidential.
Rambach cautioned against waiting for the perfect measurement approach before acting. “If you wait for clear measurement of value, you will miss a lot of opportunity,” he warned. This willingness to move forward with reasonable confidence rather than absolute certainty has enabled Schneider Electric to scale AI applications while competitors remain stuck in pilot purgatory.
AI Management Lessons for Other Enterprises
Schneider Electric’s approach to AI offers several lessons for companies seeking to scale beyond experimentation:
Start with business value, not technology. Every AI initiative at Schneider Electric begins with business needs and customer pain points, not with questions about what’s possible with the latest AI models.
Engage front-line employees in development. The people doing the work have essential domain knowledge that central AI experts lack. Effective AI requires that these perspectives be merged from the start.
Embed AI in existing workflows. Rather than asking customers or employees to adopt new stand-alone tools, Schneider Electric builds AI capabilities into the systems people are already using.
Design for scale from the beginning. Schneider Electric’s organizational model, technical infrastructure, and governance processes are all built to deploy production systems, not to create pilots.
Invest in differentiated training. Different roles require different levels and types of AI literacy. A one-size-fits-all training program won’t cultivate the capabilities needed across the organization.
Balance analytical and generative AI. Despite the current excitement around generative AI, traditional machine learning on structured data continues to deliver substantial value in many contexts.
As AI capabilities continue to evolve rapidly, Schneider Electric’s disciplined, business-driven approach provides a model for enterprises seeking to move beyond experimentation to genuine operational impact. By designing for scale, engaging front-line workers, and maintaining focus on measurable business value, the company has built an AI program that meets the objectives of both customers and employees.