Four elements—AI-ready data, context engineering, governance with LLM observability, and human oversight—were presented by Elastic and MIT Technology Review’s Insights arm as the core architecture for scaling reliable AI systems.
Gartner’s forecast that companies could abandon 60% of AI projects through 2026 without AI-ready data underpins the first priority: connected, accurate, governed data that can be retrieved in real time.
Context engineering was framed as the mechanism for feeding each query the minimum correct, current information, using tools such as RAG and vector databases to improve accuracy while limiting cost and latency.
Governance and observability were positioned as built-in controls rather than add-ons, with Elastic citing a 2026 survey showing 85% of IT decision-makers expect to enable LLM observability for internal generative AI apps.
Nearly 70% of respondents in Deloitte’s 2025 Tech Executive Survey plan to grow teams because of generative AI, reinforcing Elastic’s argument that human expertise remains essential as systems become more autonomous.
With AI agents acting as 'digital insiders,' are our current cybersecurity strategies becoming obsolete?
In the AI arms race, is building a stable foundation a winning strategy or a fatal delay?
How can we bridge the 'intention gap' to safely manage autonomous agents we don't fully understand?
Why 50% of Generative AI Projects Fail: Data Readiness, Governance, and Architectural Strategies for Sustainable AI Success
Overview
This report highlights that many AI projects fail, with at least 50% of generative AI initiatives abandoned after the proof of concept phase by the end of 2025. The main reasons are poor data quality, inadequate risk controls, escalating costs, and unclear business value. As a result, only 28% of AI use cases in infrastructure and operations fully succeed, while 20% fail outright. Poor data quality alone leads to failed AI projects and costs organizations an average of $12.9 million annually. The report emphasizes the urgent need for strong data quality, governance, and clear business alignment to achieve AI success.