A CEO's Handbook for GenAI

The Rise of Generative AI: Navigating the New Era

The release of ChatGPT in late 2022 sparked a surge of interest in generative AI. Within mere hours, users were experimenting with this novel technology, uncovering productivity hacks that quickly spread across various sectors. In the months that followed, organizations scrambled to keep up, all while navigating the unexpected challenges generative AI brought to the table. Some forward-thinking businesses have taken a more formal approach, establishing dedicated teams to explore how this technology can unlock hidden value and streamline operations.

For CEOs, however, the implications of generative AI extend far beyond immediate productivity gains. While today’s focus may center on efficiency improvements and technical hurdles, a revolution in business-model innovation is on the horizon. Just as Mosaic—the first free web browser—ushered in the internet era and transformed the way we live and work, generative AI holds the potential to disrupt virtually every industry. The promise of competitive advantage and creative destruction is real. This means leaders must move beyond initial excitement and start developing a generative AI strategy that sits squarely in the C-suite.

Addressing the CEO Challenge

This task is anything but simple. Many CEOs, often several steps removed from the technology, may feel uncertain about how to proceed. Yet the focus shouldn’t be on mastering the technology itself; rather, CEOs should prioritize understanding how generative AI will affect their organizations and industries. They need to make strategic decisions to capitalize on opportunities while mitigating risks. These decisions are built on three fundamental pillars:

• Unlocking Potential: How will widespread access to generative AI reshape innovation within the workforce?

• Redefining Roles: How will job descriptions evolve as AI redefines employee responsibilities?

• Managing Risks: How do leaders address the potential for biased or inaccurate AI outputs?

Generative AI is evolving rapidly, and while these questions are critical, there are numerous other considerations. CEOs must be ready for the moment when their existing business models become obsolete. So, how should leaders prepare for this future?

Unlocking Strategic Advantage

Generative AI has democratized access to powerful tools like ChatGPT, DALL-E 2, and Stable Diffusion, enabling users to create websites, develop advertising strategies, and produce videos with ease. This "low-code, no-code" functionality is driving adoption at an unprecedented scale.

While the initial productivity gains are significant—AI can summarize lengthy documents in seconds, reducing research time and costs—generative AI’s accessibility means that competitors have the same tools at their disposal. Simply leveraging AI for productivity won't differentiate businesses. CEOs need to identify the “golden” use cases—those that provide a true competitive edge and create significant impact.

Identifying the Right Use Cases

The key for CEOs is to uncover opportunities across the value chain that offer differentiation. This might involve driving growth through enhanced offerings, as seen with Intercom’s integration of AI into its customer engagement tools to move toward automation-first service. In sectors like biopharma, AI’s ability to accelerate research could drastically increase the value of patents by shortening time-to-market.

Once these golden use cases are identified, strategic decisions must follow. Should the company fine-tune an existing large language model (LLM) or invest in training a custom model? Both approaches have distinct advantages and trade-offs.

Fine-Tuning vs. Custom Model Training

Fine-tuning existing models can be cost-effective and help businesses hit the ground running. For example, Snorkel AI found that fine-tuning a model for complex legal classifications cost between $1,915 and $7,418, saving lawyers countless hours of work. However, this approach has limitations, particularly when it comes to relying on the core model's domain knowledge and training data.

On the other hand, training a custom model provides greater flexibility but comes with steep costs. AI21 Labs estimates that training a 1.5-billion-parameter model could cost upwards of $1.6 million. For companies seeking a unique competitive advantage, this investment might be justified, but the stakes are high.

Planning Your Investment

Timing is everything. CEOs must weigh the risks of investing too early in complex AI projects against the dangers of falling behind competitors. Today’s generative AI still has limitations, particularly regarding accuracy, and should be used cautiously in cases where variability is acceptable. Leaders will also need to consider how to fund these initiatives, whether through IT, R&D, or another budget source.

As the generative AI market evolves, research is increasingly moving behind closed doors, making it harder to keep pace with cutting-edge models. Companies that want to remain competitive will need to build strong internal teams capable of managing these advanced technologies.

Preparing the Workforce for AI-Driven Change

Much like traditional AI, generative AI is reshaping the workforce. CEOs must collaborate with leadership and HR teams to redefine roles and responsibilities as AI augments productivity. This transformation will see employees, from marketers to coders, using AI to complete first drafts of tasks, allowing them to focus on higher-level strategic work.

However, this shift must be managed carefully. Many employees fear that AI will make their jobs redundant. CEOs need to address these concerns head-on, ensuring that AI is seen as a tool to augment rather than replace human work. Transparency and effective change management will be crucial in helping employees embrace this new AI-driven environment.

Redefining Operating Models

Agile, or bionic, operating models will continue to be the most scalable. Centralized IT and R&D departments with AI expertise will be necessary to support data-driven decisions across business units. This centralization ensures that employees working with similar data have access to the same datasets, preventing siloed information from hampering AI’s potential.

In some cases, companies may benefit from establishing a senior executive role—such as a Chief AI Officer—responsible for overseeing AI initiatives. This executive would manage the intersection of business and technical requirements, ensuring that AI adoption is both efficient and secure.

Mitigating Risks in the AI Era

Generative AI presents unique risks, particularly around accuracy and security. AI models can produce inaccurate or biased results, and the potential for data breaches is significant. Companies need to establish clear policies that guide the safe use of AI and ensure that experimentation doesn’t compromise proprietary data.

Training employees on the appropriate use of generative AI is equally important. Overconfidence in AI-generated outputs can lead to serious security vulnerabilities, especially in coding. CEOs must encourage a healthy skepticism and mandate thorough review processes for all AI-generated content.

Securing the Future of Generative AI

Generative AI holds unprecedented potential, but it also challenges leaders to rethink their approach to business models, workforce dynamics, and security protocols. By crafting a thoughtful AI strategy, CEOs can ensure that their organizations remain competitive in this rapidly evolving landscape.

Leaders who are prepared to harness generative AI’s power—while safeguarding their businesses against its risks—will be well-positioned to create long-term advantage.