It’s been widely reported that most AI projects fail, and there have been many explanations offered as to why.
One of the most important factors of a project’s success is simply assembling the right mix of people. You may not realize all of the players that need to be included.
So who do you need on your team?
Let’s first outline the rules of the game. AI is a discipline of computer science that develops the ability of a machine to yield insight or perform tasks commonly associated with human intelligence. But getting systems to do these tasks takes several steps: data prep, AI modeling, simulation/testing, deployment, and ongoing monitoring/iteration are the typical stages. There are different owners for each that need to work together to get the use case up and running.
The key players are the line of business (LOB), IT administrators, AI engineers, and AI practitioners. LOB is the end-user, and AI practitioners are the developers. Many enterprises make the mistake of only including those two players in a project.
However, each of the four players above plays a critical role in implementing AI – from development and implementation to deployment. Let’s walk through their roles in the game, starting with the end-user.
LOB: The line of business is the department in the organization that is the end-user of the AI technology. They’re the ones leveraging AI to improve a workflow in their function, whether by automation of functional tasks, by expanding capacity or enhancing customer relationships. The AI application they are using needs to fit with their workflow requirements and evolve as their needs do.
Enterprise IT: The IT administrator provides the infrastructure and platform for running AI workloads. Enterprise IT must ensure that the AI workloads meet the performance needs of LOB stakeholders, while providing tools that developers are familiar with, and ensuring compliance with corporate security and privacy requirements. IT must also ensure they can effectively manage this infrastructure in a cost-effective way. According to Gartner®, IT leaders who industrialize AI solution delivery will create vastly greater value faster and compound these gains as AI solutions scale enterprise-wide*.
AI engineers: AI engineering exists within IT, as a specialized model of DevOps, overseeing the piloting and deployment life stages of AI. Gartner estimates that by 2025, the 10% of enterprises that can establish AI engineering best practices will generate at least three times more value from their AI efforts than the 90% of enterprises that do not.
AI practitioner: Last, we have the AI practitioners who create, fine-tune, and refine AI models. There are multiple stakeholders in this segment that oversee data prep, training, and inferencing life stages, and their functions can range from data scientist to application developer. These folks are the most intimately connected to AI on a day-to-day basis, and are making decisions on how to design the model for the end-user. However, their decisions also impact the AI engineers and IT administrators, so connecting everyone in the team allows for complete development and implementation at the earliest stage.
When brought together, these players help set the stage for success and create the blueprint for repeating it over and over as AI use cases proliferate.
IT decision makers can learn more about AI best practices at NVIDIA GTC, which takes place March 21-24. The free, virtual event has become one of the largest and most important technology conferences, featuring more than 900 sessions and 1,400 speakers from a who’s who of organizations including Amazon, Dell Technologies, Deloitte, Google, HPE, Mercedes-Benz, Microsoft, NASA, NFL, Pfizer, Stanford University, Visa, Walt Disney and Zoom.
Register here to learn how to harness the latest technologies to solve complex challenges.
