Maximizing the return on AI: Lessons from Google’s advertising platform

Artificial intelligence, specifically agentic AI, is quickly becoming a standard capability. Organizations across nearly every industry now have access to increasingly sophisticated models, copilots and automation tools. As these technologies become more widely available, the conversation is beginning to shift. The question is no longer whether a business has access to AI, but why some organizations using similar technologies continue to achieve dramatically different results.

Digital advertising has quietly become one of the world’s largest real-time experiments in applied artificial intelligence. Every day, platforms like Google Ads make billions of decisions about where advertisements should appear, who should see them, how budgets should be allocated and which creative is most likely to resonate with a customer. It is easy to assume that Google’s AI is responsible for those outcomes. Looking more closely at how the platform operates suggests something different.

Before Google’s AI is ever given the opportunity to optimize a campaign, the business has already made the decisions that matter most. Leaders determine whether success means increasing revenue, acquiring new customers, improving profitability, or maximizing return on advertising spend. Those strategic choices establish what the platform is trying to accomplish.

Once those objectives have been established, the Merchant Center provides the structured product and business data that enables Google’s AI to optimize toward them. Rather than serving as a simple repository of product information, the Merchant Center organizes products, pricing, availability, promotions and other business information into a format the platform can understand. Google’s AI is not asked to infer what the business sells or how its products relate to one another. That understanding has already been established through the way the business structures and maintains its data.

Seen this way, Google’s advertising platform is about far more than sophisticated algorithms. Strategy defines the objective, structured data provides the foundation and AI applies both at a speed and scale no team of marketers could replicate manually. The quality of the AI’s decisions is shaped long before the model begins making them.

The same dynamic appears well beyond digital advertising.

One recent engagement involved helping a regional U.S. bank modernize its data environment after fragmented information had limited the organization’s ability to generate meaningful insights. Although the symptoms were most visible in reporting and business intelligence, the engagement revealed that stronger analytics alone would not solve the problem. The bank first needed a trusted data foundation. Rather than beginning with new dashboards, the team modernized the enterprise data architecture, rebuilt the Customer360 model and implemented a data-as-products strategy. With that foundation in place, the organization was then able to modernize reporting and analytics, enabling more reliable insights, greater self-service capabilities and broader data-driven decision making.

From Google’s advertising platform to one of our recent engagements with a regional US bank, the conclusion is remarkably consistent. Improving the technology alone would not have changed the outcome. Better results came from strengthening the objectives, data and enterprise foundations that allowed the technology to perform effectively.

Perhaps this is where many organizations unintentionally narrow the conversation around AI. We often discuss data as though it were the end goal, when its real value lies in what it enables. AI cannot create understanding from poorly organized or incomplete information and it cannot compensate for unclear business objectives. Its effectiveness depends on the strength of the foundation supporting it.

This changes how we should think about creating value from AI. For years, organizations differentiated themselves through access to technology. Today, access to AI is rapidly becoming an expectation rather than an advantage. As the technology becomes more widely available, differentiation shifts toward the work that surrounds it.

Ironically, many of the investments that generate the greatest return from AI do not resemble AI initiatives at all. They involve clarifying business objectives before selecting a model, modernizing fragmented data environments before deploying copilots, strengthening governance so information can be trusted across the organization and redesigning processes that allow AI to scale effectively. None of those efforts generate the same excitement as launching a new AI capability, yet each one increases the value of every AI capability that follows.

Technology will continue to evolve, models will continue to improve and today’s leading capabilities will eventually become commonplace. The organizations that realize the greatest value from AI will not be distinguished simply by having access to better AI. They will be distinguished by building businesses that allow commoditised AI to create exceptional value.

The question for leaders is no longer simply where AI can be deployed, but whether the foundations are in place to turn it into meaningful value.

Learn how we help organizations turn AI ambition into impact: