Leveraging AI democratization: scaling-up innovation through accessible technologies AI models are becoming more efficient – smaller models now match the performance of their larger predecessors while being faster, easier to deploy and cheaper to scale up.27 Between November 2022 and October 2024, inference costs – the computing resources and cloud or server fees involved in training AI models – dropped 280x, hardware costs declined by 30% per year and energy efficiency improved by 40%.28 AI systems have also achieved rapid gains in sustained reasoning and workflow orchestration, achieving more with fewer parameters.29 These advances are redefining what work looks like on the factory floor. As AI shifts from automating tasks to orchestrating decisions, frontline and logistics operators are moving from “doing” to “overseeing”. Recognizing that skills in this new era evolve as quickly as the tools themselves, companies are shifting from reactive consumers to co-producers of talent, partnering with educational institutions and regional ecosystems to collaborate in defining the next generation of digital skills.30 As AI evolves, new tools are putting problem- solving power directly in the hands of frontline teams – leveraging democratized innovation to strengthen enterprise resilience. Lighthouses show how this shift draws in a new generation of digital-native talent while reinvigorating experienced employees with technologies that elevate and redefine their work.Low/no-code tools and platforms enable innovation directly from the shop floor The democratization of AI is changing both the tools of innovation and the hands that wield them. For example, low- and no-code tools and large language model (LLM)-enabled tools can reduce development costs by up to 70%, while enabling “citizen developers” (frontline workers without specialized coding knowledge) to spearhead innovation.31 The role of citizen developers is already dominant at Lighthouses – responsible for 58% of solutions in 2025.32 At SOCAR’s Carbamide plant in Sumqayit, Azerbaijan, a no-code platform empowered frontline operators without analytics expertise to optimize energy use and reduce emissions in gray urea production. Designed as a scalable, self-service product, the solution integrates a machine learning- based power and steam optimization engine with an intuitive interface, enabling rapid control logic customization by non-technical users. The platform has already been scaled up to three other sites (Figure 13).33 When operators can design and deploy their own digital solutions, agility becomes embedded in the organization’s DNA. For Lighthouses, this is what resilience looks like in practice – a workforce empowered to solve problems and incubate new ideas from the shop floor up. Global Lighthouse Network: Rewiring Operations for Resilience and Impact at Scale 22

Global Lighthouse Network 2026