1、Copyright 2025 S&P Global.All Rights Reserved.May 2025Lessons from industry leadersBest practices for delivering AI at scale451 Research Discovery ReportBest practices for delivering AI at scale| of contentsIntroduction3Methodology3Figure 1:Demographic breakdown of participants3Executive summary4Ins

2、ight 1:Prioritize scalability and recognize that the cloud is not a solve-all6Figure 2:Vast majority of organizations leveraging the cloud for AI workloads7Insight 2:Data strategy and security must be addressed at the outset of AI projects9Figure 3:Most organizations say infrastructure could better

3、support data pipeline management10Insight 3:Effective project prioritization is key to controlling costs11Figure 4:Required infrastructure upgrades for AI span security,storage,networking and compute12Insight 4:AI initiatives do not operate in isolation;organizations must emphasize both integration

4、and collaboration14Figure 5:Wide array of stakeholders involved in infrastructure decisions to support AI/ML workloads15Implications:Laying the foundation for AI at scale16AI strategy16Networking16Security16About the author17Best practices for delivering AI at scale| are actively investing in AI as

5、the technology rapidly advances,but not all are reaping the benefits.According to 451 Researchs Voice of the Enterprise:AI&Machine Learning,Use Cases 2025 survey,84%of organizations reported active investment in generative AI.However,46%of these respondents did not see significant positive impacts f

6、rom those investments against any organizational objectives assessed in the survey.All are working to build the scaffolding that they hope will bring them AI value,but a clear divide is emerging between organizations effectively leveraging AI and those that are not.This new report on best practices