With the explosive growth of AI and ML-driven processes, companies are under more pressure than ever to drive innovation. For many, adding a Data Science capability into their organization is the easy part. Deploying models into a large, complex enterprise that is operating at scale is new, unchartered territory and quickly becoming the technology challenge of this decade. This article takes an in-depth look at the primary organizational barriers that have not only hindered success but often prevented organizations from moving beyond just experimentation. These obstacles include challenges with fragmented and siloed enterprise data, rigid legacy systems that cannot easily be infused with AI processes, and insufficient skills needed for data science, engineering and the emerging field of AI-ops. Operationalizing AI is hard, especially at the fast pace at which the business operates today. This paper uses real-world examples to guide a discussion around each of these hurdles and will equip industry leaders with a better understanding of how to overcome the challenges they will face as they navigate their data and AI journey.

随着AI和ML驱动流程的爆炸性增长，公司承受着前所未有的更大压力来推动创新。对于许多人来说，在他们的组织中添加数据科学功能是很容易的部分。将模型部署到规模庞大的大型复杂企业中是一个新的，未知的领域，并迅速成为本十年的技术挑战。本文深入研究了主要的组织障碍，这些障碍不仅阻碍了成功，而且经常阻止组织超越仅仅进行试验。这些障碍包括零散和孤立的企业数据，无法轻松注入AI流程的刚性旧系统以及数据科学，工程学和AI-ops新兴领域所需的技能不足等挑战。AI的运作非常困难，尤其是在当今业务快速发展的今天。本文使用真实的示例来指导围绕这些障碍的讨论，并将使行业领导者更好地了解如何克服他们在导航数据和AI旅程时将面临的挑战。