Site Change story Top five use cases Impact Haitian Flavouring & Food Foshan, People’s Republic of ChinaTo maintain its cost advantage and produce soy sauces with consistent flavour, Haitian implemented digital transformation to address rising consumer demands for diverse products and manage increasingly complex orders. The site deployed over 50 digital use cases, nearly half of which were powered by AI, to optimize operations. Despite a 54% increase in SKUs and a 64% rise in small-batch orders, these efforts reduced raw material waste by 33.6%, cut product defect rates by 39.1% and shortened order lead times by 38.7%. ML-enabled enabled raw material yield optimization 23.2% Raw material waste Smart nose-enabled soy sauce aroma quality improvement 25.3% Sensory defect rate Digital process quality assurance with near-infrared spectroscopy 95.9%Average changeover time Neural network based closed loop adjustment for precise filling 49.4% Volume deviation APS with genetic algorithm to enable high-mix low- volume production 38.7% Order lead time Guizhou Tyre Guiyang, People’s Republic of ChinaGuizhou Tyre faced the challenge of meeting highly customized orders (over 60% of all orders) for heavy load operations, which required strict design, verification and reliability standards. To speed up product design and ensure consistent and stable output, the site implemented over 40 4IR solutions to improve agility and productivity in manufacturing. These solutions, which included AI-enabled design, advanced data analytics and flexible automation, resulted in a 57% reduction in defects, a 68% increase in labour productivity and a 34% reduction in inventory levels.AI-enabled performance simulation and formulation recommendation 66%Customized product design cycle One-click extrusion changeover and parameter adjustment 11% OEE Vision-based high-resolution X-ray inspection for tyres 99% Defect leakage LLM-powered digital assistant for maintenance and quality control 35% Defect rate AI-enabled demand forecasting and S&OP coordination 33% Forecast accuracy Aramco North Ghawar, Saudi ArabiaNorth Ghawar Oil Producing Complex’s history began with the start of Saudi oil production in 1938, and today manages assets dispersed over a 12,500 km2 area. To meet increasing demand while reducing both operational costs and emissions, the complex launched a 4IR strategy. It pioneered oil producing intelligence, boosted existing asset reliability and upskilled young talent by leveraging over 65 solutions, including advanced analytics, AI-powered digital twins and Aramco large language model (LLM) GenAI. As a result, oil production increased by 8.44% while scope 1 and 2 emissions reduced by 8.21% per barrel of oil equivalent.Prescriptive analytics at the edge for autonomous well operation to maximize oil production and optimize manpower 8.44% Throughput Digital twin integrated planner to schedule multi- facility production with ML- based power optimization 17.8% Energy intensity ML-based flaring prevention model to minimize process upsets towards net-zero GHG ambition 35%Flare-causing process upsets AI-based asset failure prediction with LLM capability to extend asset life & expedite recovery 77% OEE Facility operator co-pilot powered by GenAI to empower crew and assure throughput availability 75% Training cost Global Lighthouse Network: The Mindset Shifts Driving Impact and Scale in Digital Transformation 41

Global Lighthouse Network 2025