Putting AI to Work for Marketing


Today, all of my conversations increasingly revolve around AI, specifically how generative AI is going to change marketing, how we work, how we create, and how we deliver exceptional experiences for customers. Marketing has come a long way from decades of mass advertising and untargeted direct mail. AI and technology have helped marketers be more precise as to who we reach out to, with what message, and offers that we feel are most relevant.

We’ve moved towards customization, but personalization continues to be a future state goal. We’re trying to create moments when marketing helps us identify exactly what we need, or maybe discover the one thing we didn’t even know we wanted. And that’s where generative AI comes in, with large language models and multimodal models that have the ability to make use of structured, unstructured, unlabeled data across internal and external sources to discover patterns, correlations, and insights that we couldn’t see before, and to help us gain insights and take action on those discoveries. And we don’t yet know everything that’s going to come out of this huge moment of generative AI disruption, but for marketers, Gen.AI is already making its mark in two key ways: content creation and personalization. The first wave of generative AI content creation is happening right now, with marketers exploring and using apps to generate marketing blogs, web banners, posters, and so much more. Generative AI is unleashing the power of advanced creative tools for companies both large and small, with the ability to engage in the creative process through simple text prompts that provide guidance to the models as well as flexibility to iterate and regenerate at the press of a button. The creative process for large enterprises, though, requires us to step back and recognize the need for the generative models that we use to have a clear and concise understanding of the brand, the company values, the product portfolio, and even the legal considerations. For generative AI to be used at scale within the enterprise, tuning and training models must be brand-knowledgeable. Once we have models that understand your company, your brand, your products and services, and most importantly, your customers, how are we going to take advantage of this powerful capability?

Let’s start at the beginning of the creative process and change the way creatives ideate, provide guidance as to the customer’s need, persona, and concept, and put GenAI to work creating a raft of ideas to jumpstart your product teams, writers, and designers. Once you get past ideation, you can land on your idea and move into what I’ll call the creative production process. The magic of a big creative idea needs to be brought to life through a series of disciplined steps. We call this the content supply chain. And for enterprise clients, it’s truly like a production line. Once you’ve landed on your target parameters, your message, your offer, and your creative have to be put into assets that can then be executed across channels. And all those steps are incredibly time-consuming, and a lot of them are very manual. I need to take this ad. I’ve got to put it into 16 different languages in five different formats. Executing all these derivative versions of your creative is a production task that can be automated with confidence.

That’s where generative AI can play a big role because it can take those tasks and effectively move them through the production process a lot faster. And that’s going to free up marketers from doing very mundane and routine activities to allow them to put their focus back on strategic and creative work. Now the second disruptive opportunity is personalization, where generative capabilities may have even more impact. For years, AI has been helping marketers identify targets, predict attrition, recommend products, and more, but the AI couldn’t create content. Generative AI and its ability to create or customize at speed mean it’s now possible to personalize messages in much more granular ways. This allows for micro-segmentation, addressing very specific needs and attributes while still ensuring a brand’s voice and offer are properly represented. With the power to generate and create in near real time, it opens the door to much more relevant, perhaps even individual, messages. and we call this personalization at scale. Imagine a system that can predict, read, and react to customer inputs on the fly, responding to and engaging with customers. Now let me share an example. I’m a marketer for a chain of movie theaters, and I see a bunch of complaints coming in. I want to be able to respond very quickly and even zero in on appropriate responses to retain my customers and the shared wallet. So let’s say a customer named Stephanie complains that the projector broke during the movie she’s watching. Generative AI can read her complaint, summarize it back, and generate a response very quickly. Stephanie, I’m so sorry the projector broke down yesterday at Queensway Mall. And as a valued customer, your next movie is free. GenAI can assemble the information from her viewing history and the incident she experienced and provide a highly personal response.

Now that’s just one example. But you can imagine that same type of scenario playing out across almost any point of customer contact. What we know about customers and what we can infer from data about prospective customers can let us create a really engaging experience in near real time. But on a practical level, how do we get started? In our recent global survey, 67% of CMOs stated that they plan on implementing generative AI in the next 12 months and 86% in 24 months. The writing is on the wall, so to speak, but there’s still a lot of uncertainty as marketing leaders look for ways to incorporate AI, but to do it safely, responsibly, and effectively. The complexity of implementation is the number one concern. Enterprises are naturally cautious. I get it. However, this movement toward generative AI is driven not just by the big tech firms but by the customers themselves. They’re interacting with GEN.AI solutions now in their everyday lives, and their expectations of highly personal interactions are setting a new bar for companies. 76% of CMOs we surveyed indicated they must implement GEN.AI or lose their competitive edge in the market. You want to put some thought into what you can effectively buy and use off the shelf, versus what you should curate yourself, and how a blend of those things might look in your organization. There will be generative AI baked into a ton of products, and we’re already seeing it. This is AI at a pervasive level. Think about GenAI built into your word processors and your image editors. It’s a business accelerator, but it’s not a differentiator. Everybody will have access. Then, at the next level, you’ll have AI as a service with marketing-centric offerings and API integrations available. You’ll start to see some separation here as you get cleverer about how and where you implement it. but you may only have limited control over the model or the data. And ultimately, if the service is available to you, it’s available to your competitors too. And finally, you have the platform approach, where you have the power and capability to customize and oversee your own AI. You can pick and update models, you can train and tune with your proprietary data, and you can build the specific solutions you need when you need them. This is your AI laboratory. To experiment with all the tools and infrastructure, you need to get creative and start inventing. To really differentiate, you need to get to the platform level. That’s where your proprietary data becomes an extraordinary source of value. It holds the key to everything that makes your business your business. And for marketers, it truly reflects your brand. Data is everything when it comes to AI. There’s no AI without data, so treat it like a precious resource.

For marketers, the foundation of any GEN.AI program is ensuring that your model has a deep understanding of your brand. To do this, you need to collect, digitize, and connect the data that reflects these elements. That’s what you’ll use to tune, train, and refine your models. Legal compliance with local laws is something we’re all going to need to consider as countries define how they’re going to protect consumers and customers in this new world of generative capabilities. And since the data you leverage for customization is sensitive, you’ll need to ensure security and governance are built into your processes. Now you can start doing all this right now, even before you have GEN.AI in place. One powerful yet challenging aspect of GEN.AI models is that they’re continually learning. So it’s important to build out a program that continually refreshes and monitors your models with access rules and evaluation to ensure they’re on brand and not demonstrating bias. You might think of this as your Gen. AI model steward. It’s all part of a critical AI governance practice. I talked earlier with a focus on customer-facing applications, but we shouldn’t overlook the value of internal employee applications. There’s an opportunity to accelerate productivity, but beyond that, internal transformations create an invaluable opportunity for learning and experimentation. Your marketers will experience organizational change management to embrace and harness these powerful new tools. This is a human and technological change we’re experiencing. Just think how generative AI can be an active assistant, you know, riding along with marketers to help them ideate, create, and innovate. This might mean streamlining processes, aligning tasks with roles, and building real-time dynamic training. Just think how GenAI can be a game changer for market research. All that unstructured data you’ve collected may hold customer or product insights that you may not have even known to look for, but they’re visible with GenAI and accessible to marketers through simple, natural language queries.

The sky is the limit here. There’s real value in being covered internally. And as you build AI and data skills in this environment, you’ll be better prepared to create external applications that wow your customers. Generative AI is going to change the way marketers work, and almost every activity will be affected in some way. But that doesn’t mean that everything has to change at once. Your company is going to build up your AI marketing capability the same way you built up your security expertise and your data management. Find one great use case and tackle that. As you build your Gen.AI muscles, you’re not alone. We’re all learning and re-imagining together, and there’s help available if you need it.

About Anushka Agrawal

Leave a Reply

Your email address will not be published. Required fields are marked *