The AI Engine

D

ata is the fuel that powers today’s digital landscape. As more organizations look to meet ambitious AI objectives, they’ll need to build the infrastructure and data foundation to help support these grand plans. 

The good news? Organizations that are aligning their models with well-organized data structures are already reaping the benefits:

“Customizing AI models to fit specific business objectives unlocks measurable value and fuels innovation,” said Tony Giordano, Senior Partner Data and Technology Transformation and Data Services Global Leader at IBM Consulting.

To witness this approach in action, look no further than the US Tennis Association (USTA) which partnered with IBM to develop generative AI models that use tennis data to create insights and original content for the US Open website and app. The USTA used IBM watsonx and data models to inform the large language model. The models were then trained to translate tennis data into cogent descriptions, including match reports. 

Still, many businesses struggle to scale up AI. According to Giordano, a robust and well-planned strategy can help overcome obstacles like data silos, which can prevent AI systems from accessing all the data sources needed for a meaningful analysis. 

Leveraging data

Of course, successfully aligning AI solutions with business goals requires strong leadership. Organizations should involve internal stakeholders — particularly those closest to data and business processes — to guide strategic AI customization efforts. If clients fail to define clear objectives or carefully manage their data, they may face unnecessary costs and slower time-to-value, Giordano noted.

As AI workloads grow more complex, flexible, and scalable, finding the right data environment becomes an important consideration. Hybrid cloud strategies combine cloud-based infrastructure with on-premises systems, allowing businesses to enhance data integration. They also help avoid the costs of data silos, such as duplicated time and effort managing data in disconnected environments. 

Hybrid cloud systems don’t always share data seamlessly, so organizations need a structured approach, such as centralized data governance frameworks, to help ensure smooth data transfer between multiple environments. 

The right structure simplifies the process of managing diverse data sources, making it easier to harness data for AI-driven insights and innovation. Tony Giordano, Senior Partner Data and Technology Transformation and Data Services Global Leader, IBM Consulting

This approach fosters interconnected data ecosystems and supports AI deployments that require real-time access to different data sources. It’s no surprise, then, that two-thirds of companies surveyed in “AI in Action 2024” are using or planning to use a hybrid cloud to run their AI workloads.

“Without a hybrid cloud strategy, it can be difficult to realize the potential of AI,” Giordano added. 

Adopting multimodel approaches

As AI evolves, organizations are increasingly adopting multimodel, platform-based approaches to help solve complex challenges. Instead of relying on a single AI model, they deploy several specialized models optimized for different use cases. That allows for greater flexibility and more precise outcomes.

The larger the AI model, the more computing resources it requires, which can mean increased latency, cost, and energy consumption. That’s why organizations are moving toward customizing smaller models with data around their own users and domains. Maryam Ashoori, Director of Product Management, watsonx.ai

Establishing technological foundations

To deploy AI successfully, organizations also require a robust infrastructure that can scale with evolving business needs. The foundation must be closely tied to the organization’s strategic vision to ensure AI efforts have the greatest impact.

This involves modern data architectures and technologies like hybrid cloud to ensure data consistency and availability across the enterprise. Companies can build this foundation in phases, beginning with scalable, flexible infrastructure like hybrid cloud and data lakes that can grow alongside their AI initiatives.

“Capabilities of modern AI models can be used to work better with data,” Stephan Bloehdorn, Partner and Practice Leader of AI, Analytics & Automation at IBM Consulting, said. 

To support their AI initiatives, businesses must prioritize seamless data access, security, and governance. By establishing a solid technological framework today, companies can prepare to adapt to future developments in AI, helping scale up their AI solutions and maintain a competitive edge.

Organizations that succeed with AI will be those that take a comprehensive, business-driven approach by combining robust data strategies, customized AI models, and a strong technological foundation. It’s also important to have internal champions for AI who can ensure that initiatives align with the company’s strategic goals and receive support throughout the organization. By investing in the right infrastructure, embracing hybrid cloud strategies, and adopting a multimodel approach to AI, businesses can thrive in an increasingly AI-driven world.

Learn more about how IBM can help you develop a robust technological foundation to unlock the benefits of AI.

This post was created by Insider Studios with IBM.