Skip to main content
Exploring what’s next in tech – Insights, information, and ideas for today’s IT and business leaders

New research explains why some AI projects succeed while others never get off the ground

How enterprises are learning to get the most of AI/ML.

Artificial intelligence seems to be a fickle technology. Why do some businesses succeed while others struggle to realize value from AI and machine learning? And how does technology investment impact a successful AI strategy? Hewlett Packard Enterprise recently commissioned Emerald Research Group to find out the answers.

The Emerald Research study groups organizations into three categories, based on their level of AI adoption: developing, using, and advanced. Characteristics of each group are very much in parallel with the level of capabilities these organizations show in their digital transformation journey. For example, advanced organizations (represented by 26 percent of respondents) are using AI either in ways that have disrupted traditional business models or as a primary method of generating business value, which is key to digital transformation (see Figure 1).

These advanced organizations are bringing the edge and AI together to create the intelligent edge, enabling them to outperform their competitors, while organizations that wait to build an AI strategy and invest in edge solutions risk falling behind.

Barriers to success

In the research, organizations told us their biggest challenge is finding the talent and expertise to implement their AI strategies. As organizations grow in AI sophistication, they often quickly realize they face a critical skills gap and lack the expertise needed to implement their strategies.

In addition, organizations have difficulty understanding the ROI of AI. The ability to articulate the return from an investment in AI improves as organizations mature in their AI sophistication. This struggle to quantify and measure ROI is almost universal across developing, using, and advanced organizations.

As organizations classified as developing and using strive to realize their AI strategies, they tend to struggle with having quick and easy access to data. Moreover, many are skeptical about the veracity of the data they work with. Organizations wishing to build a successful strategy earlier in their AI journeys must first address these data requirements by adopting a data-first approach.

Secure, flexible AI is best

The more advanced an organization is with AI and ML, the more advantage it sees in using modern cloud architecture and tools, including as-a-service delivery, for such demanding applications.

Moving AI and advanced analytics workloads to the cloud or an AI-as-a-service (AIaaS) model is not a trivial task, especially when critical business processes and data are concerned. But organizations expect the benefits to outweigh the costs and risks once their transformation to a cloud-first AI/ML model is complete. Along the way, they will face opportunities, challenges, and decisions that must be made. The proper framework provides a structure and common language to understand where they are in their journey. It also allows organizations to benchmark their work against best practices in the industry and correctly prioritize their next steps methodically.

The research also tells us that deployment flexibility, having an end-to-end platform, and security are organizations' top three criteria in selecting an AIaaS solution. As organizations grow and mature in their level of AI sophistication, they look to AIaaS to maintain flexibility in deploying their solutions in the cloud or on premises (see Figure 2). This finding supports what we see with the AI and data client projects HPE experts deliver.

Scaling AI solutions and making them enterprise ready requires continuous collaboration between teams and business units via a flexible, open, secure platform. The ultimate goal is that everyone can benefit from a shared pool of organizational talent, AI models, tools, and data, wherever they are.

This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.