Three roles CEOs need to play to scale Generative AI

CEOs belong on the front line of decision-making—especially when it comes to imagining and implementing GenAI for business success. They need to set the vision, communicate that vision, and then make the right investments to accelerate the journey toward success.

The strategic opportunities GenAI presents require CEOs to take a deep dive into their organization’s technology agenda. Given the staggering choices and outsized impacts of GenAI, CEOs should be at the centre of key decisions they would normally delegate. They’ll need to decide whether to be a first mover or a fast follower, whether AI is needed for innovation or productivity, and whether they should be building or buying AI capabilities.

The buck stops with the CEO

Yesterday’s white-hot innovations can become today’s headaches. Take multicloud for example. Many enterprises rushed into piecemeal cloud agreements without establishing a central decision-making hub, and later found themselves with technical sprawl in desperate need of streamlining. Without appropriate caution and consideration, GenAI adoption could take the exact same path.

As more players roll out AI options, CEOs can learn from the lessons of the past and take time upfront to make key decisions. They can start by asking:

  • Should we focus our investments on a few key choices or should we maximize optionality while the competitive market for GenAI plays out?
  • How can our organization build flexibility in our execution approach?
  • How will we measure direct and indirect cost and performance implications?
  • How do we embed trust and guardrails in AI model development?

In the context of GenAI, CEOs can apply their experience as deal-makers, venture capitalists who bet on winning strategies, and champions of business priorities in three important areas: securing computing power, selecting an ecosystem for their large language models (LLMs), and establishing centres of excellence.

Secure computing power access

With the widespread adoption of GenAI, the necessity for swift model training and execution has emerged as a critical business requirement. Conventional compute infrastructure relies on central processing units (CPUs) that handle data sequentially. However, when dealing with LLMs, specialized AI chips are needed to enable massively parallel processing, a crucial element for efficiently processing terabytes of data through algorithms based on GenAI. Research predicts that the market for specialized chips will be well over US$60 billion in 2024 and climb up to US$120 billion by 2027.1

Some organizations’ needs may be met by niche cloud providers who specialize in GPUs. However, CEOs with the AI ambitions of pursuing innovation and competitive advantage may want to secure more robust computing power. Many such CEOs are engaged in conversations with GenAI hardware companies and are negotiating with chip makers to secure the right level of resources.

As deal-makers, CEOs must consider investor sentiments and engage with their executive leadership team, specifically their chief information officer (CIO) and chief technology officer (CTO), to facilitate alignment between hardware procurement strategies and overarching business objectives. Additionally, emerging technologies such as edge computing present new opportunities for decentralized AI processing. By staying abreast of technological advancements and market trends, CEOs can make informed decisions to future-proof their organization's computing infrastructure.

Select an AI ecosystem

The value of GenAI hinges upon the data it consumes. This presents a problem for CEOs, as most LLMs available are not built with out-of-the-box domain, industry, or organization-level specificity. Private LLMs can deliver clear advantages in choice, cost, and control, while enabling enterprises to retain their intellectual property. To capture value and scale, enterprises need to select the LLMs and broader AI ecosystems that suit their specific needs.

CEOs can view their organization’s data as a bargaining chip. This untapped data trove could be highly valuable to companies building AI products. CEOs can consider valuing current and future-state data assets as potential inputs to new models, as long as they concurrently secure or anonymize the data to avoid trust and regulatory concerns.

As with securing hardware, CEOs need not delve too deep into the technical specifics of different AI models. Instead, they must calibrate their company’s AI ambitions and determine whether it will be a first mover in building custom LLMs or buy them later. Adopting a venture-capitalist-like mindset, the CEO can leverage their understanding of the marketplace and established relationships with major players to determine which bets can be made safely, while considering their broader portfolio and being thoughtful about every round of investment.

Establish AI centres of excellence

A recent Deloitte and Fortune CEO survey (Winter 2024 Fortune/Deloitte CEO Survey Insights) found 80% of organizations are already implementing or likely to implement GenAI to accelerate innovation, while 96% are doing so to increase efficiencies. To reap these expected benefits, organizations are establishing AI centres of excellence (COEs). These hubs serve as catalysts for innovation, enabling organizations to conceptualize, develop, and deploy AI solutions at scale. A COE can bring together a cross-functional group of AI experts and interest holders to focus organizational efforts and create a consistent approach to governance and guardrails (e.g., by hiring ethicists).

The CEO’s role in implementing a centre of excellence is to champion its establishment and provide the support and resources needed to get it launched. This includes ensuring critical elements such as hardware, data needs, and governance are sufficiently funded.

CEOs must also engage in winning the hearts and minds of all interest holders, including customers, employees, the board, and society at large. Working with their chief legal or risk officer, CEOs can help ensure that all members of the enterprise feel prepared for what’s on the horizon.

While many employees are fearful of AI transformation, they are also eager to understand what future jobs and skills may look like. CEOs should foster a culture of AI fluency and innovation and champion a clear purpose for AI adoption, as a matter of supercharging people (not replacing them).

By leveraging AI centres of excellence as platforms for knowledge exchange and collaboration, CEOs can position their organization as a front-runner in the race to gain the Generative AI advantage.

Ever-expanding CEO priorities

The AI revolution will continue to alter the roles and responsibilities of CEOs in the years to come. Already, they need to be more tech-savvy than ever, given how important technology is to competitive advantage and ways of working. As AI becomes even more embedded into knowledge work, the keys to AI adoption are expanding out of the tech leader’s domain to become CEO priorities.

This is especially true while AI adoption is still nascent. CEOs can bring their experience to bear on the many macro and micro decisions that need to be made when a technology is both very new and very impactful. As we continue our series on leading an AI-fuelled organization, we’ll delve into more aspects of the roles CEOs must play toprepare their organizations to pivot to the Generative AI future.

You can read the full report via our global site.

1 Gartner, “Gartner Forecasts Worldwide AI Chips Revenue to Reach $53 Billion in 2023,” August 22, 2023.

Did you find this useful?