The accelerating integration of Artificial Intelligence (AI) within corporate sustainability practices has generated new challenges in understanding how autonomous, agentic systems contribute to environmental accountability. This study investigates how four leading technology firms—Apple, Microsoft, Google, and Meta—incorporate Agentic AI into their sustainability and ESG reporting between 2021 and 2024. Employing comparative content analysis, the research develops and applies a detailed codebook for identifying agentic characteristics such as autonomy, adaptivity, goal-directed optimization, coordination, and embeddedness in physical infrastructures. The findings reveal substantial variation among firms: Microsoft demonstrates the most advanced integration, particularly in energy management and emissions tracking; Google shows strong implementation in data center optimization and renewable energy forecasting; Apple’s approach is selective and operationally focused, while Meta remains in an exploration phase. Building on these insights, the study proposes a practical framework of AI Sustainability Impact Indicators to support standardized assessment of AI-related environmental impacts, including metrics for energy share, lifecycle emissions, renewables usage, and Responsible-AI alignment. The research contributes to bridging the gap between AI governance and sustainability disclosure, offering an empirically grounded foundation for future policy and corporate accountability frameworks.