The 3 Things AI Actually Does in Marketing: Predictive, Generative and Agentic AI

Over the past few years, the landscape of technology has shifted dramatically. Utilizing AI in marketing has moved from a novelty to an absolute necessity for leaders. According to McKinsey’s “The State of AI in 2025” report (published March 2025), 88% of surveyed organizations now report regular artificial intelligence use in at least one business function, compared with 78% in the prior year’s survey.¹ Navigating marketing in an AI-driven world requires more than just trying new applications. To provide clarity, Associate Professor Roy Wollen—creator of the “Marketing in the Age of AI” course—and Associate Professor and Director of the IMC Professional program Chris Cahill shared their perspectives during a spring 2025 webinar on the subject.

This post explores their three-part framework, real brand examples, ethical considerations, the future of marketing teams and how to future-proof your career.

It’s Not About the Tools—It’s a Marketing Operating System

Imagine the CEO stops you in the hallway and asks about your AI strategy. Responding with a favorite application, like ChatGPT or Claude, is the wrong answer. The right answer is an AI marketing roadmap tied directly to business outcomes, such as customer acquisition, retention, growth and advocacy.

In the current environment, Wollen and Cahill argue that artificial intelligence functions as an AI marketing operating system—redesigning workflows to achieve significant business impact. Many teams are shifting away from tool-operator mindsets; the speakers suggest we are now in the age of agent management.² Tool choices must be governed by clearly defined outcomes rather than underlying mechanics.²

What Is AI in Marketing? Predictive, Generative and Agentic AI Explained

At its core, artificial intelligence refers to systems capable of processing information and learning to perform tasks. However, navigating predictive versus generative versus agentic AI means understanding that, in this framework, AI is grouped into three functions in a marketing context.

Predictive AI

Predictive AI marketing focuses on forecasting customer behavior. Marketing managers rely on these models to analyze data, derive insights, understand market opportunities and develop strategies for acquiring and retaining customers.³ This includes identifying the next best experience (the most relevant interaction to serve a customer at a given moment), assessing churn risk, modeling the propensity to buy and calculating lifetime value. For example, a streaming service might use predictive models to identify subscribers likely to cancel and trigger a retention offer.

Generative AI

Generative AI in marketing centers on creating new experiences in the moment as users interact. This technology generates highly relevant messages with bespoke tone, imagery, and copy at high volume.⁴ Marketers use these systems to produce personalized chat interactions, video content and customized offers in real time. For instance, a retail brand might generate thousands of personalized email variations tailored to individual browsing histories.

Agentic AI

Agentic AI marketing goes beyond content creation by executing actions autonomously to achieve higher-level goals.⁵ One key use case involves automating and scaling what is already working, though agents can also explore new approaches and learn from results. For example, an agentic system can handle autonomous campaign optimization, allocate bids and budgets and conduct test-and-learn initiatives at scale. A travel company, for instance, might deploy an agent that continuously reallocates ad spend across channels based on real-time booking data.

These three types of technology commonly work together in sequence, though teams often iterate and loop between them. Predictive models identify who to target and when, generative systems create the tailored assets for those targets and agentic systems deploy and optimize the materials automatically.

How Is AI Being Used in Marketing? Real-World Brand Examples

The following examples highlight how AI-driven marketing functions in the real world.

Retail Personalization/‘Clienteling’

Anticipatory experiences rely on recognizing customer identity across channels, whether through account-based identification or biometric systems. Amazon previously utilized its Amazon One palm-recognition service for entry, identification and payment at hundreds of retail locations.⁶ The biometric system demonstrated how physical stores could connect in-store associate “clienteling” with seamless transaction flows. According to GeekWire, the company shut down its palm-payment system at physical retail locations in early 2026 after limited customer adoption.⁷ For marketers, this illustrates the potential of identity-layer technology to enable real-time personalized service at the point of sale.

Augmented Reality Shopping

The Home Depot’s Project Color application uses augmented reality to help customers visualize room changes.⁸ The application’s “See It in Your Space” feature can recognize lighting conditions and paint around other objects to provide a highly realistic preview.⁸ This bridges the gap between digital visualization and physical execution, showing marketers how immersive tools can reduce purchase hesitation.

B2B Buying Signals

In the business-to-business (B2B) sector, roughly 40% of accounts show meaningful signs of buying activity at any given moment.⁹ Platforms that capture these signals help teams detect when a dormant account is back in market. When an account shows a spike in intent, the system can trigger routing and notify sales representatives immediately.¹⁰ This matters for marketers because it enables precise timing of outreach, reducing wasted effort on unready prospects.

Why Personalization Alone No Longer Wins

Seamless, anticipatory experiences are now table stakes rather than differentiators. Today, 80% of customers say the experience a company provides is just as important as its products and services.¹¹ Customers expect brands to adapt to their changing needs, and 73% expect better personalization as technology advances.¹¹

Brands that rely on traditional demographic personas instead of individualized, real-time predictions may find it harder to compete as customer expectations rise. Currently, 84% of marketers confess to running generic campaigns, even as consumers demand instant, conversational interactions.¹² Missing basic personalization expectations creates an immediate competitive disadvantage.

Navigating Challenges in AI Adoption: Ethics, Governance and Search

As organizations adopt these systems, AI marketing ethics must remain a priority. More than half of organizations using these models have experienced negative consequences, with inaccurate outputs being the most common issue.¹ Human-in-the-loop oversight is critical.

Strong governance structures—including formal policies, review processes, and ongoing monitoring—uphold safety throughout the system’s lifecycle.¹³ As a risk-management best practice, compliance processes should be automated but not bypassed, and extra caution is warranted in regulated industries such as healthcare and financial services.¹³

SEO vs. AEO vs. GEO

Search behavior is shifting, creating a new challenge for marketers evaluating SEO, AEO and GEO strategies. Customers are increasingly getting answers directly from artificial intelligence rather than clicking through traditional search engine results.¹² According to Gartner’s “Integrate AEO and SEO” research (2025), a significant share of web content will be created primarily for consumption by search and AI systems by 2026.¹⁴ Marketers must distinguish between traditional Search Engine Optimization (SEO), Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO) to maintain visibility. Use SEO for traditional organic rankings, AEO for direct-answer platforms such as voice assistants and GEO to ensure content is reliably cited by AI-generated responses.

Will AI Shrink or Grow Marketing Teams? Two Expert Views

The question of whether AI will replace marketing jobs remains debated among professionals. Survey data shows that 32% of respondents expect artificial intelligence to decrease workforce sizes, while 13% expect an increase.¹ Wollen and Cahill offer differing perspectives on this trend.

Professor Wollen’s view

Wollen suggests that internal teams and agencies will shrink as execution becomes heavily automated. As marketing automation AI takes over routine tasks, fewer people will be required to do higher-leverage work.

“What’s going to happen right beneath our feet is marketing teams are going to shrink, and I think that there’ll be fewer people doing more impactful work. And media buying is going to be autonomous and optimized using AI. So, fewer hands, fewer people, but more efficiency and better decisions—more strategic [work],” he said.

Professor Cahill’s view

Cahill argues that automation creates more opportunity and output rather than fewer jobs. He compares the current shift to the transition from manually written tables of numbers to the adoption of Lotus 1-2-3 and Microsoft Excel, noting that better tools historically expand a department’s capabilities and drive demand for skilled professionals.

“The advent of tools [like Lotus and Microsoft Excel] didn’t change the number of people operating inside of financial services institutions,” Cahill noted. “What it allowed people to do was not spend so much time focused on writing down numbers and doing sums by hand. [Instead] it opened up new avenues to go deeper, to build out much more complex financial models. I expect that the marketing suite is going to be similarly transformed by AI.”

Both experts agree that cross-functional fluency between data, analytics, marketing and creative teams is an essential skill moving forward.

Preparing for an AI-Driven Marketing Future

To stay competitive, professionals must build a strategy that prioritizes gaps in their workflows. A recommended first step is to conduct an audit of your current technology stack to assess readiness.² From there, you should deconstruct workflows into discrete tasks to identify what can be automated. Finally, establish clear guardrails and permissions before giving autonomous systems meaningful responsibility.²

Additionally, marketers should increasingly treat tool fluency as a baseline requirement rather than a unique skill. Resources such as Northwestern University’s Knight Lab offer ongoing research into media and technology to help professionals stay current.

The strategic skills needed to lead AI-driven marketing efforts are exactly what advanced programs are designed to develop.

Master AI in Marketing with Medill’s IMC Professional Program

Applying artificial intelligence effectively is a leadership and strategy skill, not just a technical one. The live online Master of Science in Integrated Marketing Communications (IMC) Professional program at Northwestern prepares you to lead in this evolving landscape. The program is designed to help you develop the strategic frameworks that can support your organization’s growth and your career advancement.

You can strategically tailor your education through the curriculum, which includes the Marketing in the Age of AI course as well as other options such as Digital Marketing Activation, Digital Marketing, Media and Innovation, and Managing Digital Products and Technologies. The program also features immersive global courses to deepen your expertise. Earning your degree can help you build a powerful network and navigate the intersection of data and creativity, elevating your career to the next level.

Ready to take the next step? Schedule a call with our admissions team by reaching out to an admissions outreach advisor today.