How to Deliver Successful AI Projects by Reducing Risk and Boosting Performance

How to Deliver Successful AI Projects by Reducing Risk and Boosting Performance

By Mike Leone, ESG Senior Analyst


Let’s be clear: Almost all enterprises—including a lot of midsize and even smaller ones—hope to exploit the business potential of artificial intelligence (AI) and machine learning (ML). And why wouldn’t they? AI and ML have already been used by many organizations to deliver real value, from reducing stock-outs in retail to improving production quality and reliability in the manufacturing of pharmaceuticals. Those and other initial successes do two things: Encourage others to step up their AI efforts, and signal to competitors that they need to go big on AI or risk losing out to more aggressive competitors. But despite the collective excitement over the potential for artificial intelligence, many enterprises continue to struggle to overcome challenges in turning pilot programs into successful deployments with tangible business value. An end-to-end, AI-ready platform can reduce risk, improve the likelihood of a successful deployment, and generate demonstrable returns.

The Focus on AI

With the necessity to adopt AI or be left behind, it’s no surprise that spending on AI solutions is spiking. ESG research notes that 62% of organizations plan to increase their spending on AI over the next year. This spending covers a wide range of investment areas, including hardware, software tools, IT services, application development, deployment services, hiring AI-proficient staff, and modernizing their workflows and business processes to embrace and leverage AI/ML.

62% of organizations plan to increase their year-over-year spending on AI, including investments in people, process, and technology.

There’s just one thing: Successful AI deployments are difficult to accomplish today. Technical and business challenges can drive a wide range of complexities, which means organizations need to really do their homework. Specifically, they need to pay close attention to planning system design, coming up with the right development and deployment methodologies, selecting the right infrastructure and tools, ensuring data cleanliness, centering on the right metrics to measure progress, and partnering with the right technology experts.
There already is ample evidence that a large portion of AI pilots and sandbox projects never make it to the production stage, having been hamstrung by one or several issues. These challenges are exacerbated by the fact that not many in-house IT professionals have all the requisite skills and experiences to develop, deploy, and run a successful AI initiative.

82% of organizations have seen value from their AI initiatives in as little as six months.

The good news, however, is that organizations that make the right investments in technology, processes, and people can see demonstrable results with significant impact to the business from their AI initiatives. According to ESG research, 82% of organizations have seen value from their AI initiatives in as little as six months. ESG also points out that organizations have seen a wide range of significant benefits from their AI programs, such as improved quality, enhanced cybersecurity, identification of new market opportunities, and faster tactical response to unique customer requirements.

Acknowledging and Overcoming AI Deployment Challenges

Succeeding with AI programs is difficult. There are many challenges and obstacles that organizations need to overcome to even come close to their intended outcomes. Finding, developing, and deploying the right strategy puts big demands on organizations to get it right, or they run the risk of project delays, cost overruns, and unfulfilled expectations that can alienate decision-makers to AI. Additionally, organizations are taking a hard look at their existing infrastructure and recognizing that they must modernize their technology stack to support the high performance, security, and scalability requirements of resource-intensive AI use cases.
More specifically, organizations must understand, acknowledge, and overcome these challenges:
The AI skills gap: Even with the rapidly growing popularity of AI, organizations continue to struggle to find the right talent—both inside and outside the organization—to handle the technology, processes, and business justification for appropriate investments.
Increased risk due to disparate technologies: Pinpointing and dealing with widening sources of risk, such as integrating disparate technologies and tools with legacy infrastructure.
• High-performance requirements: Achieving the essential high performance for data-intensive AI workloads, especially deep learning analytics.
• Complex software launch cycles: The demands of staged production cycles, from sandboxes and proofs of concept to enterprise-wide production systems tightly integrated with a wide range of systems, applications, and business processes.
• Managing complex infrastructure: There are multiple areas of stress in managing the existing infrastructure stack including resource sharing, integrated development environments, GPU processing, CPU processing, and data storage. Without a plan to deal with these challenges, it becomes increasingly difficult to attain AI proficiency without help.
• Enterprise-wide deployment: Challenges in operationalizing AI throughout the enterprise. ESG research highlights that AI deployment is the AI lifecycle phase that generates the most headaches for organizations.
Finally, organizations need to come to grips with the reality that they typically will have AI workloads all over the map—on-premises, in the cloud, and back and forth, often in different environments for different phases of a project. This makes managing deployments and ongoing production systems extremely challenging, making the selection of the right solutions partners essential.
AI and ML Workload Deployment Environments


Private Cloud

Hybrid Cloud


Keys to Creating and Delivering High-value AI Solutions – Purpose Built AI Platform

No one must be told that there are tons of options when it comes to potential technology suppliers for AI infrastructure and tools. However, there are far fewer suppliers that have actual experience in developing and deploying the kind of infrastructure solutions essential for powering today’s AI workloads, as well as the track record for a technology roadmap that scales to meet organizations’ future needs.
Instead of simply creating a laundry list of suppliers for hardware, software, and services that could be used to build an AI solution, organizations need to prioritize the potential for those suppliers to become true partners in the planning and execution of systems for AI workloads that make a difference.
Organizations should demand that potential technology partners can deliver infrastructure and software solutions purpose-built for AI. Those specific infrastructure demands should include:
• Hybrid and multi-cloud capability: Organizations need a clear understanding of how to facilitate fast, reliable, and secure workload migration from on-premises to the cloud and back, and from cloud to cloud.
• Optimize hardware infrastructure deployments: Ability to optimize hardware infrastructure utilization to avoid over-provisioning hardware in processing, storage, and network capacity.
• GPU integration: Seamless integration of powerful compute accelerators that meet the ever-increasing performance needs of AI workloads.
• AI infrastructure expertise: Proven expertise in successful, rapid, and repeatable deployment and provisioning of AI systems.
• Virtualization and container support: Organizations also should pay close attention to technology partners that can support virtualization, a critical and efficient approach to AI systems for increased ability and to accelerate the creation of true AI-powered environments. ESG research indicates that 38% of organizations currently utilize containers to support AI and ML workloads and another 33% currently utilize virtual machines; this is key to organizations’ ability to experiment and test hypotheses faster and more cost-efficiently.

Organizations' preferred virtualization and container approach to support AI and ML workloads:

38% currently utilize containers
33% currently utilize virtual machines

Leveraging the VMware and NVIDIA Collaboration

The demands of AI-powered solutions that are purpose-built for demanding AI workloads typically transcend the capabilities of any single supplier, regardless of their resources and experience. For these kinds of solutions, it is often advantageous to seek out suppliers that have created a formal collaboration to deliver AI infrastructure solutions from end to end in the data center, on the edge, and in the cloud—both physical and, increasingly, virtual.
VMware and NVIDIA, each market leaders in their respective technology sectors, have pooled their skills and resources to develop an enterprise platform for AI workloads. The transformational and strategic nature of AI means that these projects garner a lot of attention, visibility, and scrutiny inside organizations. Both individual workgroups that will benefit from AI and the board rooms where executive governance on IT investments is paramount care deeply about these initiatives. This makes it paramount to work with technology suppliers that not only bring skills and proven capabilities to the table individually, but also have a demonstrated ability to work synergistically for a customer’s benefit.
Many Fortune 500 organizations already have worked with both VMware—the global leader in multi-cloud infrastructure solutions—and NVIDIA—the market leader in accelerated computing for AI and computer graphics. Together, these market leaders have built an AI-ready platform for enterprise-wide AI projects across a wide range of use cases.
The solution is constructed on VMware vSphere running the NVIDIA AI Enterprise software suite. NVIDIA AI Enterprise is a cloud-native platform comprising of AI and data science software that is purpose-designed and optimized to run on VMware vSphere on NVIDIA-certified systems. It is engineered for AI workloads typically running in modern hybrid cloud environments, which have fast become a standard architecture for IT workloads, in general, and for AI workloads specifically. The platform is jointly certified by NVIDIA and VMware with enterprise support and is performance-optimized for data-hungry AI workloads for AI training and inference.
Figure 1. NVIDIA AI Enterprise Suite

Source: VMware

The NVIDIA AI Enterprise Suite is optimized for vSphere, the enterprise workload platform for traditional and next-generation applications. vSphere supercharges performance with DPU and GPU-based acceleration, enhances operational efficiency through the Cloud Console, seamlessly integrates with add-on hybrid cloud services, and accelerates innovation with an enterprise-ready Kubernetes environment that runs containers alongside VMs.
Figure 2. VMware + NVIDIA AI-Ready Enterprise Platform

Source: VMware

The platform runs on NVIDIA-Certified Systems, which are accelerated servers validated to run demanding workloads such as AI training and inference, and data analytics. ConnectX Smart NICs and NVIDIA BlueField data processing units (DPUs), provide organizations with software-defined infrastructure for high-performance networking and enhanced security. VMware and NVIDIA AI-Ready Enterprise Platform delivers near-bare-metal performance across multiple nodes, streamlining the time to production of new AI applications and services.
Figure 3. Near-bare-metal Performance Across Nodes

Source: VMware

The Bigger Truth

For organizations to reap the benefits of “first movers” in the deployment and leveraging of AI, moving quickly is essential. Achieving speed to results means organizations must combine the right technology, the right strategy, the right business metrics, and the right resources to avoid falling behind.
Organizations should adopt an AI-ready enterprise platform that can accommodate the kinds of high-performance, high-availability, and ultra-secure workloads where AI is ideally positioned. Specifically, organizations should select an AI-ready platform—and, of course, the right AI-proficient technology partners—designed not only for today’s needs but especially for the coming generation of AI-powered use cases.
VMware and NVIDIA have teamed up to develop an end-to-end enterprise platform designed to leverage AI for transformative workloads—the kinds of solutions that separate winners from also-rans and encourage organizations to take even bigger leaps with AI tools, technologies, and services. Together, VMware and NVIDIA help organizations reduce complexity and risk, increase deployment speed, and accelerate and scale “science projects” to a truly differentiated solution in production environments.

This ESG White Paper was commissioned by VMware and is distributed under license from TechTarget, Inc.

All product names, logos, brands, and trademarks are the property of their respective owners. Information contained in this publication has been obtained by sources TechTarget, Inc. considers to be reliable but is not warranted by TechTarget, Inc. This publication may contain opinions of TechTarget, Inc., which are subject to change. This publication may include forecasts, projections, and other predictive statements that represent TechTarget, Inc.’s assumptions and expectations in light of currently available information. These forecasts are based on industry trends and involve variables and uncertainties. Consequently, TechTarget, Inc. makes no warranty as to the accuracy of specific forecasts, projections or predictive statements contained herein.

This publication is copyrighted by TechTarget, Inc. Any reproduction or redistribution of this publication, in whole or in part, whether in hard-copy format, electronically, or otherwise to persons not authorized to receive it, without the express consent of TechTarget, Inc., is in violation of U.S. copyright law and will be subject to an action for civil damages and, if applicable, criminal prosecution. Should you have any questions, please contact Client Relations at

Enterprise Strategy Group | Getting to the Bigger Truth™

Enterprise Strategy Group is an IT analyst, research, validation, and strategy firm that provides market intelligence and actionable insight to the global IT community.