Gartner預測到2026年60%的AI專案將因缺乏這4大架構支柱而失敗
gartner預測到2026年由於數據準備不足，60%的ai專案將被廢棄，強調可靠的數據管道和上下文工程對於從原型轉向可擴展的生產級ai系統至關重要。
gartner predicts that organizations will abandon 60% of ai projects by 2026 if they lack ai-ready data, according to an mit technology review insights report in partnership with elastic. the report identifies four enduring architectural pillars necessary for scaling ai: data quality, context engineering, governance, and human expertise. it emphasizes that while models evolve, the foundation must support reliable retrieval and decision-making, highlighting that context engineering—often more critical than prompt engineering—ensures models receive precise, relevant inputs to maintain speed and accuracy. successful architectures require real-time, unified data foundations and vector databases to support advanced techniques like retrieval augmented generation (rag) as companies move toward agentic systems.