Start experimenting. To experience the potential of this technology first-hand, senior leaders must dive in and explore its capabilities. CMOs should encourage their teams to identify valuable applications, experiment with models, and start building transformative use cases. One approach is to create cross-functional agile marketing pods that can take on a task, such as launching a marketing campaign, with as much GenAI as possible. Once marketers find ways to hack their processes, the organization’s data scientists and engineers can automate and build connections to enhance it—for example, by using enterprise versions of LLMs or building application layers to produce output in more usable forms.

Seek game-changing outcomes. CMOs should aim to achieve step-change gains in productivity through innovative and disruptive approaches. Doing so creates a different risk-reward calculus for setting priorities. CMOs need to identify “golden” use cases that enable them to use their core data and IP assets uniquely to create a competitive edge. Training the models on IP and fine-tuning them with key data (marketing performance as well as consumer, brand, and market research) will also ensure that the outcomes are sufficiently differentiated from what competitors can produce. Settling for small gains with big ROIs may seem attractive, but a company can’t afford to walk when its competitors are running.

Establish an enterprise-wide model. Scale and competitive advantage are elusive in the absence of the right solution and architecture. Today’s GenAI model market is volatile, which exposes companies to two extremes: either selecting an unsuitable enterprise-wide provider or having to cobble together a collection of providers. Consumers and end-users have fueled unprecedented growth of LLMs, but now tech companies are developing suites of enterprise solutions to spur greater innovation. Developers can select different models that meet their needs from a library of LLMs, the goal being to find LLM providers that complement their existing cloud or tech supplier while retaining flexibility on the last-mile applications on top (providers of bots, content creation, and so on). These tools will enable CMOs and their teams to improve process efficiency, personalize customer interactions, inspire innovation through unconventional creativity, and create customer value in new ways.

Implement responsible AI guidelines. If an organization prohibits GenAI use or lacks centralized guardrails, one of two things is probably happening. Either employees are using GenAI anyway—professionally or privately—and recognizing opportunities for productivity improvements or the organization is falling behind competitors that are already pursuing and may be achieving double-digit-percentage gains in productivity.

The balance lies in incentivizing experimentation with GenAI while mitigating the numerous risks. Using AI responsibly means developing and operating AI systems that align with organizational values and widely accepted ethical standards while also achieving transformative business impact.

Most CMOs in our survey see AI regulations as inevitable, and the vast majority have undertaken some form of self-regulation. (See Exhibit 4.)