How AI Agents Are Transforming Marketing Strategy

Dan Loney: Not only are we seeing AI used to find out things about our work or our studies, but it's increasingly being used in the world of marketing. The goal is to better connect with the consumer. But with the growth of large language models, do companies also need to better connect with the public? Especially bots because they are, in many cases, making the connection with humans.

Pleasure to be joined once again by Stefano Puntoni, marketing professor here at the Wharton School. He recently delivered a presentation to Wharton's Executive Education course on this topic. Stefano is also co-director of Wharton Human-AI Research. Stefano, how much are we seeing humans rely on bots when they're making these decisions, like when you're talking about in retail?

Stefano Puntoni: There's a lot of exciting change happening in marketing. You mentioned retail. Certainly, that is one very important domain. There are others. We were convening a group of executives as part of a new program on AI in marketing that Eric Bradlow, our vice dean for AI Analytics, and myself were co-leading. And a lot of the conversations were in this space.

I argue that we are basically undergoing two revolutions in marketing. The first one is well underway. And that revolution started in 2022, when ChatGPT was released. That revolution is about search and discovery. How consumers find information about the marketplace. People increasingly learn about the market through interactions with chatbots like ChatGPT and others.

The second revolution is just starting out. This year is turning out to be the takeoff year for personal agents. We'll talk about that later. But essentially, AI agents are becoming more common. They're still adopted by only a small minority. But we can see that trajectory moving very fast. And that means that it's no longer about using LLM-powered AI agents for search and discovery but actually delegating decision-making authority on behalf of consumers that are informing different steps on the customer journey. In that sense, we are seeing increasingly the buying decision being co-created by humans and algorithms.

Loney: Because you have so many people using ChatGPT right now, and I think we will have more people use it in the years ahead, how should that technology impact marketing?

Puntoni: If you look at the recent paper that the team of economists at OpenAI, led by Ronnie Chatterji, published on what people actually do on ChatGPT, one of the three largest uses was search and discovery. But one of the main elements of that process will have to do with a marketplace in some shape or form. People trying to solve a problem, people looking for a product to do a particular job, people learning about the product category and a lot of other similar things. Where in the past, you might have gone on a search engine like Google and then mined through a bunch of different links to try to explore the available offerings in a category, which brands offer which attributes at which level, for which price points, what's the availability of these products, and everything else.

Now, they might do that more efficiently using chatbots, where you can ask a chatbot to tell you about that ski jacket versus that bicycle versus that whatever. And it will tell you which brand, for which attributes you're interested in, then even help you design a comparison matrix with price points and tradeoffs and helping you make a decision.

Loney: Take us into the world of the AI agent and the bigger role that it's playing right now.

Puntoni: AI agents are basically systems. They're not just one thing. They are systems powered by language models that also integrate a lot of other things, including the use of APIs and scripts and other things that basically support decision-making. They are big in professional contexts, obviously. But they're also becoming big in more consumer-facing functions.

This year we've had OpenClaw and Clawbots in multiple discussions where these open-source agents have found very wide adoption in, let's say, tech-savvy markets. Then you've had Perplexity, increasing revenues very quickly on the back of the release of AI agents. You've seen, for example, Anthropic releasing lots of plugins to the Claude chatbot to give it increasingly agentic features. There are a lot of other examples of that sort, which show how AI agents are, after about a year of promise where everybody was saying they're going to come, they're actually here.

Loney: How does this impact firms and their approach in that relationship they have with the consumer, with the LLMs as the vehicle in between?

Puntoni: Certainly, there are many ways in which these revolutions are going to impact the way the company is thinking about their marketing or about the way they interface with external environment. Some of them are quite obvious, already apparent today. Others are unclear. We know things are going to change, but how exactly and for whom, there is some uncertainty there.

But for the more certain part, we know that companies need to think very carefully about their visibility to language models and bots generally, so that when you type a query, on such on ChatGPT or when you ask your Clawbot to do a search for you, it will find you as a brand and make it an available option to the consumer. Because obviously, if LLMs ignore you, then you're not even there as part of the conversation.

For example, many companies sell their products across many different platforms, many different e-commerce websites. And for many organizations, the back end is a bit of a mess, meaning that the same product might be described with different text tags or even different codes depending on which platform it is being sold on.

Now, when these bots send out search queries across the internet to gather information to support a consumer who wants something, they will use the authority of the brand or the product with a very strong queue for deciding whether to recommend it. Authority can be gleaned by a lot of different sources, including customer reviews, including brand reputation and awards of different shapes. But one important thing is consistency across platforms, when the bot is searching for the same products in a lot of different places.

So, if your back end is a mess and the same product is appearing as different products on every platform, the bot will not find you. They will think that you are selling a very diffuse set of stockkeeping units, and that has little authority, when in fact it is always the same product. It is just that the bot doesn't realize. This is one concrete example, which is really about getting your IT infrastructure, product coding, and interface with the vendors systems in a way that supports bot decision-making.

Another one that is quite obvious is also that your website — your own web-facing real estate —  has to be amenable to bot search. The internet that we have today is the internet that was  designed for people, not for bots. It is highly visual. It supports the way that we like to do things. You've got these drop-down menus that make search and navigation easier. But those are hard for bots. We need to think about what we can do to make it easy for bots to find the information that they need on our websites.

You’ve started seeing the emergence of protocols and standards, and companies starting to release pages that are not designed for people, but are designed for machines. That is going to be a big area of investment and improvement too.

Loney: Don't companies have to also figure out a way to integrate the two of them together?

Puntoni: Absolutely. We will need basically two internets, one for people and one for machines. And these two internets need to somehow speak the same language, meaning that there has to be consistency in the product information. There have been things like availability and prices and product attributes and all the rest of it.

We don't quite know exactly how to do that, but there are lots of startups and lots of efforts in this space. I expect a lot of investments and interesting innovation happening at this boundary between marketing, retail, and supply chain.

Loney: In the presentation you made to this Exec Ed class, you talked about a topic that is very much in and around AI in general, and that's that of bias. Where does bias factor into this discussion around AI?

Puntoni: The moment that we delegate decisions to agents, then it becomes important to understand, how do they make decisions? We know that there is bias emerging in language model output. Bias can mean different things depending on the discipline. If you are a statistician, bias is deviation from validity, so it's basically a formula. If you are more into culture and society, bias might be about discrimination. And if you are a psychologist, bias might be about cognitive heuristics and biases that get us to make suboptimal decisions. All of those are relevant in the context of bots.

I'm a behavioral scientist, so basically I'm an applied psychologist trying to understand how our thinking is shaping consumer behavior and the behavior of other individuals in the market. I'm especially interested in the cognitive bias aspect. It's a very weird idea where, as a psychologist, the question you might start to ask if you are interested in marketplace is not only how people think but also how machines think.

Loney: One of the things that you also mentioned is a human component is still important in all of this.

Puntoni: Yeah, absolutely. The human component is important in many ways. One of them is the consumer will want to retain control over that process to the extent that they feel that control is needed. You're going to have a delegation process where consumers are going to learn to trust agents, and then we're going to delegate to more and more authority as the agent demonstrates that they're doing a good job. There is that process of human-AI interaction.

But you also have potentially another process, which is that of thinking about what is the role of human labor? Now, if you are interacting with companies today, often you interact through a chatbot, and that can be convenient and effective. But when the default shifts away from interacting with a human service provider to interacting with an AI system, then the question becomes: How would you deploy human labor to have maximum impact in the market?

Sometimes you might find that companies can deploy humans to really signify care for the relationship, the importance of the customer that we care. So, companies should think strategically not only about the deployment of AI labor, but also the deployment of human labor. Because I think doing so could really create more valuable relationships for companies.

Loney: Do you hope that there's an overall message from this presentation?

Puntoni: My big-picture takeaway of all of this is that it's not going to be human alone, it's not going to be AI alone. It's going to be human and AI. And companies that are smart about capitalizing on the collective intelligence of the organizations — both its AI systems and its human experts — are going to be the ones that are going to thrive.