Research shows that an inadequate or lack of purpose-built infrastructure capabilities is often the cause of AI projects failing.
As the use of artificial intelligence (AI) becomes more mainstream and used in more aspects of daily operations, businesses start out by relying on compute infrastructures traditionally employed for HPC. While that approach offers some help, increasingly, what businesses find is that they need an infrastructure that is purpose-built for their AI workloads. That was the findings of a new IDC study.
Specifically, IDC recently debuted its AI InfrastructureView, a deep-dive benchmarking study on infrastructure and infrastructure-as-a-service adoption trends for artificial intelligence and machine learning (AI/ML) use cases. The global survey will be run annually and will include 2,000 IT decision-makers, line of business executives, and IT professionals, the majority of which influence purchasing of AI infrastructure, services, systems, platforms, and technologies.
The survey results show that while AI/ML initiatives are steadily gaining traction, with 31% of respondents saying they now have AI workloads in production, most enterprises are still in an experimentation, evaluation/test, or prototyping phase. Of the 31% with AI in production, only one-third claim to have reached a mature state of adoption wherein the entire organization benefits from an enterprise-wide AI strategy. For organizations investing in AI, improving customer satisfaction, automating decision making, and automating repetitive tasks are the top three stated organization-wide benefits.