How to pick the right hyperscaler for enterprise AI

Insights

• Adopting enterprise AI is intricate, but hyperscalers can simplify and accelerate the process by streamlining key factors.
• Choosing pre-packaged solutions from hyperscalers helps organizations implement AI faster, more affordably, and with greater ease.
• But, the effectiveness of hyperscalers depends on organizations creating optimal conditions for seamless integration.
• It requires robust technology infrastructure for plug-and-play solutions and a clear strategy for hiring or upskilling talent.

Implementing enterprise AI can be a complex journey for organizations, requiring careful coordination of multiple factors to ensure success. Hyperscalers can reduce the time and effort needed for this process. However, for them to deliver maximum value, organizations need to establish the right conditions. This includes optimizing their data infrastructure, adopting integration-ready technologies, and equipping employees with the skills to effectively use the new tools. These foundational steps can pave the way for a smoother and more impactful AI implementation.

Why pick a hyperscaler

Leveraging AI technology can help organizations with activities from cataloging their products and determining price elasticity of demand for them to forecasting demand and supply as part of supply chain planning. While most organizations have the will to jump on the AI bandwagon to achieve their goals, as per Infosys’s Enterprise AI Readiness Radar, only 2% of respondents were AI ready across all five readiness dimensions identified — talent, strategy, data, governance, and technology. One of the major challenges companies face in implementing AI is the costs associated, or getting business value from the implementation due to not selecting optimal use cases.

Opting for pre-packaged solutions that come with hyperscalers such as Oracle, SAP, and ServiceNow is helpful for organizations as it makes AI implementation faster, cheaper, and easier. Hyperscalers come with advanced pre-trained models, APIs, and development frameworks, eliminating the need for organizations to build AI infrastructure from the ground up. The plug-and-play nature of the solutions reduces implementation time by automating processes and minimizing the need for manual effort. It also allows flexibility and scalability in terms of the solutions being integrated with the organization’s existing systems. This approach not only reduces effort but also cuts costs by avoiding the trial-and-error expenses of developing in-house AI solutions.

“The Infosys Oracle solution empowers the technology partner to guide clients through their AI adoption journey. Due to the embedded AI aspect of the Oracle AI-first solution, which means that Oracle has been continuously adding capabilities of AI within the solution, the client can avail of the most recent benefits,” says Ashish Kumar, associate vice president and North American sales head at Infosys’s Oracle practice. “The solution’s extensibility uses AI services from OpenAI, and extends the platform capabilities of Oracle for customized use cases for the client, simplifying implementation.”

While hyperscalers can help organizations address most of their pain points related to AI implementation, organizations also must pick the right technology partner to deploy the hyperscaler solution. That said, how well the hyperscalers can work with them depends on organizations creating the right conditions for seamless integration. This includes establishing a foundational technology infrastructure that supports plug-and-play solutions, and developing a clear plan to hire the right talent or upskill existing employees.

How to set the stage for hyperscalers

The journey starts with selecting a technology partner or vendor who comes from a functional domain in an industry that matches the organization’s needs. For example, providing logistics solutions for a consumer goods company, or manufacturing for a retail company. The partner should not only provide technical and strategic support but also help in scaling the solution over time. Collaborating with a trusted partner helps in smooth working. “Organizations should pick a vendor/partner who's flexible to their demands, and has partnerships with multiple hyperscalers, not being tied to one, and can recommend the right hyperscalers to the organization,” says Srikanth Sripathi, associate vice president at Infosys’s Oracle practice.

He emphasizes that the next step for the organization is to transition to a cloud platform, which is to move from on-premises data and software to a software-as-a-service (SaaS) model. “This enables the creation of a centralized data lake, consolidates all their data in one place, and significantly simplifies the AI implementation process. Additionally, it addresses their AI licensing needs, eliminating the need for extra investments in licenses, as the Oracle solution includes embedded AI capabilities and options for extensible AI — further enhancing their readiness for adoption.”

While the technology partner will assist in identifying high-value use cases that deliver ROI to the organization, only those in the organization itself will fully understand the nuances of which use cases have the potential to be impactful. Creating a basic list of such use cases in advance allows the organization to guide the technology partner to narrow down the cases easily.

Needs of teams across geographies can differ based on regional regulations or customer requirements. As the cost of using a hyperscaler is significant, it becomes crucial to maximize economies of scale by picking a platform that addresses the needs of the entire organization. Involving cross-functional and cross-geographical input ensures the solution is comprehensive, avoids redundancy in services, and eliminates inefficiencies such as duplicate pricing negotiations.

“Organizations should establish a cross-departmental and cross-location network of stakeholders, supported by a structured cadence to align on the requirements and priorities for the hyperscaler solution. This collaborative approach ensures that key aspects are collectively agreed upon, and architectural decisions are signed off with a unified consensus. These decisions cannot be made in silos by a single business unit or location. Gaining buy-in from the entire organization on the platform from an architecture perspective, across geographically dispersed teams, is essential,” says Ashish Kumar.

Choosing the right hyperscaler is just the start, however: Organizations must be ready to implement the technology to get maximum value from it. Research from the Business Talent Group shows that 71% of organizations struggle with leveraging AI effectively due to a shortage of in-house expertise. Infosys has also found that lack of skills, knowledge, or resources is one of the primary generative AI-related challenges companies face. Organizations should put learning programs in place for employees before the deployment so that they can start using it straight away. Additionally, organizations must be ready with other skills, including data science.

The right talent is an essential component of a successful AI deployment — and that doesn’t necessarily mean losing jobs, or indeed a big hiring budget. Says Srikanth Sripathi: “For a leading semiconductor design services provider, we maintained that we do not recommend job losses because of AI automation, and that we can help them scale with the same set of people. Even if they were growing three-fold, they would not need to hire three times more the resources. The hiring would be minimal.”

Organizations must bring their employees with them: It is not enough just to train them up. They must also emphasize that the technology is being used to amplify human potential, and allay fears of job losses.

Platforms like Oracle can play a critical role by offering scalable, secure, and user-friendly solutions to implement enterprise AI, tailored to diverse business needs. By leveraging such platforms and working with suitable technology partners, companies can accelerate their AI adoption and drive innovation.