In the Age of AI, Marketing Boils Down to Being a “Moroccan Carpet Seller”
Today's narrator
 CEO Seiichiro Fujiwara After marketing financial products and establishing a venture capital firm, he founded eLife Inc. 1999. Hobbies include marine sports and snow sports. Dog lover. |
Hello, I am Fujiwara, CEO of eLife Inc.
Our company has now completely shifted the focus of our business to AI, and we refer to ourselves as an “AI company.” Personally, I use and compare several AI systems on a daily basis, including OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Gemini.
They all use the same technology—large language models—and are built upon the same body of research. Nevertheless, when conversing with each of them, there are differences in the content of their responses, their choice of words, and their way of thinking. I sense a certain kind of personality or human-like quality that makes me want to call it “the ChatGPT-ness,” “the Claude-ness,” or “the Gemini-ness.”
This is likely due to a variety of factors, such as the data used for training, human adjustments, and the policies of the development company. The creators’ philosophy and culture may also be reflected in the AI’s behavior.
Now, we’re a company that has focused primarily on marketing. So what we want to consider here is how the existence of AI like this will change the future of marketing.
Is AIO Really the Heart of Marketing?
In communities where marketers with a strong interest in AI gather, “AIO” is currently the hot topic. There isn’t a standardized term for it yet—it’s sometimes called GEO or AEO—but simply put, it’s the AI version of SEO.
When consumers ask an AI, “I want a product like this,” how can a company ensure its products are included in the response? How can it get the AI to reference its website and information? This approach is about optimizing content and data to achieve those goals.
Of course, it’s necessary to organize product information so that AI can read it accurately. I’m not saying that AI is pointless.
However, I’ve never been a big fan of SEO. And for the same reason, I feel a little uneasy about AIO as well.
It’s because it feels like they’re trying to steer search results and AI responses in a direction that suits them. Even if we analyze the algorithm’s tendencies and adjust our wording accordingly, if the underlying algorithm changes, we’ll just end up floundering all over again.
Rather than that, shouldn't we be thinking about how the advent of AI will change consumer behavior itself?
AI Acts as an Intermediary Between Businesses and Consumers
Basically, I think consumers will start asking AI what they should buy in the future. In some areas, that shift has already begun.
With ChatGPT, users can specify criteria to search for products, compare multiple items, and narrow down their options to find the ones that best suit their needs. A system is also being developed in which companies provide product feeds, and the AI uses information such as price, inventory, and reviews to suggest products.
At present, purchases are generally made through a company’s website or app, but as AI agents become more capable in the future, it would not be surprising if AI were to handle the entire process—from searching for and comparing products to making a purchase.
In other words, AI will act as an intermediary between companies and consumers. I’d like to call this “AI-intermediated Marketing”—or, in Japanese, “AI-mediated marketing.”
In traditional marketing—including advertising—the focus has been on how to influence the end consumer. What kind of ad will elicit the response we expect? Which celebrity should we feature to pique people’s interest? What kind of website will people find appealing? Naturally, surveys and other research have also been conducted with people in mind.
However, if AI is introduced between companies and consumers, marketing will need to adopt a “human-to-AI” perspective in addition to the “human-to-human” one.
Creating a visually appealing website or featuring a popular celebrity doesn’t immediately render these efforts meaningless. However, that alone is unlikely to be a reason for AI to recommend that product to someone. This is because AI places greater emphasis on specific information—such as the product’s features, who uses it and how, and the results they’ve achieved.
In the future, data designed for AI to read—as opposed to websites intended for human viewers—may become the primary point of contact connecting companies and markets. If we imagine an even more extreme future, websites might lose even the premise that they are “meant to be viewed by people” and instead become places filled with code and data designed for machines to read.
AI Will Come to "Know a Person Better Than They Know Themselves"
There’s another major change: AI has started to remember its users better.
ChatGPT and Claude have a memory feature that retains information about users—such as their personal details, preferences, and work-related context—from past conversations. I’ve even found myself looking at my own memory and thinking, “Did I really talk about something like that?” Of course, users can review, edit, or delete the content. Still, there’s no doubt that the more you use these tools, the more the AI comes to understand you on a deeper level.
Hobbies, preferences, family structure, work, lifestyle, what a person has bought in the past, and how they rated those purchases. As this information accumulates, AI may eventually come to understand a person better than they even realize themselves.
If wearable devices—such as earbuds—become even more widespread in the future, it’s possible that data on our activities 24 hours a day, 365 days a year will be accumulated. If that happens, AI won’t even need to wait for us to ask questions.
"It's about time to replace this product."
"Given your current lifestyle, this product would be a good fit for you."
In this way, AI will begin to guide consumers toward making purchases.
When it comes to expensive items like cars and homes—or things a person is particularly particular about—people may continue to agonize over them and make their own decisions right up until the very end. However, when it comes to everyday items and consumables, I think more and more people will simply accept AI recommendations almost as is.
VOC is what you need to have AI make the selection
So, what data should companies collect to enable AI—which has a deep understanding of each individual consumer—to effectively match them with the company’s products and services?
Ultimately, I believe that the “Voice of the Customer” (hereinafter referred to as VOC)—the feedback from people who have actually used the product—is what matters most.
Product information includes facts such as size, capacity, materials, and features. These are, of course, necessary. However, product information alone does not reveal how that product actually functioned in a person’s daily life.
How did it work for families with children? Was the amount enough for someone living alone to use up? Was it easy to use on busy mornings? How did it work for people with sensitive skin? Reviews, posts on Social media, and conversations within the community all include these specific real-life contexts.
One of the tasks AI excels at is identifying highly relevant information from large volumes of data. If AI knows a consumer’s lifestyle and preferences, it can find VOC from someone in a similar situation and use that as a basis for recommending products.
In traditional marketing, consumers were broadly grouped by factors such as age and gender and viewed as “clusters.” However, with AI-mediated marketing, it becomes possible to connect the vast array of VOC with the vast array of consumers on a one-to-one basis.
The important thing here is not to generate a large number of comments that are simply favorable to the company.
For example, even a negative review stating, “The scent was too strong,” can serve as a reason to buy for people who like strong scents, while providing useful information to help those sensitive to scents avoid a bad purchase. What’s needed for highly accurate AI-driven matching isn’t just positive feedback, but diverse and authentic real-life experiences.
That is why companies need to create an environment where as diverse a group of people as possible can experience their products and services, and where those experiences can be documented.
At long last, the voices that have been gathered within the community are coming to life
If you want to call it a coincidence, then I guess it is.
However, the “community” that we at eLifehave been building since our founding was designed precisely for that purpose: to create a space where companies and consumers can engage with each other on an ongoing basis and to gather the VOC generated there. That is exactly the kind of work we have been doing all along.
However, up until now, we’ve felt frustrated about how to make the most of the VOC we’ve taken the trouble to collect.
No matter how many thousands or tens of thousands of comments there were, company representatives simply couldn’t read them all in detail. Even when trying to use them for product development or advertising copy, they ultimately had no choice but to group similar opinions together and treat them as a few clusters or “representative voices.”
Sites like Amazon already offer features that use AI to summarize large numbers of reviews. While this is a very convenient feature, summarization also involves averaging out disparate opinions. A unique experience shared by just one person—one that isn’t part of the majority—may be lost in the summarization process.
However, once AI is able to identify VOC tailored to each individual consumer, there will be no need to force an average. This is because it can analyze the voices of tens of thousands of people at a speed far surpassing that of humans and pinpoint a single experience that is meaningful to that individual.
One-on-one marketing between sellers and buyers has been my long-held ideal. With the advent of AI, I feel like what I’ve always wanted to do is finally becoming a reality.
In the sense that what we’ve been doing so far is exactly what it means to live in the age of AI, we may have been very lucky.
Into the World of Moroccan Carpet Sellers
It was quite a while ago, but I once went to Morocco to buy a rug.
I walked into a store filled with carpets and started looking around to see which one to buy, when the shopkeeper said, “Come on, come on—why not have a cup of mint tea first?”
While sipping mint tea, we talk about what kinds of things we like and where we’d use them. Then the shop owner says he won’t sell any rugs that day and tells us, “Come back tomorrow.”
The next day, when I visited the store again, there were different rugs on display than the day before. As I looked at them and chatted with the shopkeeper, he said, “Well, see you tomorrow.” This went on for several days, and after about a week, I was finally able to buy a rug that I truly wanted.
The shop owner didn't try to talk me into buying the first thing he saw; instead, he took the time to understand my preferences through conversation and helped me find the perfect item for me from his vast selection.
I’ve always believed that this kind of interaction is the very essence of marketing—and perhaps the ideal shopping experience.
Ever since the days when TV commercials featuring celebrities singing and dancing were common and the term “GRP” was used, the advent of the Internet has gradually brought us closer to that world. And with the advent of AI, we may finally be able to achieve something quite similar to what a Moroccan carpet salesman does—tailored to each individual consumer.
Who Will Take Control of the "Carpet Seller"?
On the other hand, there are also concerns.
The question is whether we should entrust the entire process of matching businesses with consumers to a small number of giant AI companies.
If AI comes to understand a person better than the consumer does and even decides what they should buy, that would create an enormous amount of power. Is that AI truly selecting products for the consumer’s benefit? Or is it choosing products that benefit the company operating the AI? It may become difficult for consumers to tell the difference.
It’s possible that in the future, AI agents will compare products and even complete purchases without the user’s knowledge. While this is convenient, it’s also a little scary.
To prevent a single platform from monopolizing everything, it is becoming extremely important for both individuals and companies to determine how to protect and build “their own space.”
This is not a problem that can be solved by laws and international rules alone. Ultimately, I believe we have no choice but to secure our own territory through technology.
For example, on the corporate side, rather than entrusting everything to the public cloud, approaches are already emerging that involve running LLMs in local environments isolated from the outside world or building RAG systems that reference proprietary data (VOC) within closed environments. For individuals, solutions are likely to emerge that allow them to separate their behavioral history and preference data from major AI providers and manage them securely on their own devices.
Furthermore, rather than allowing AI agents to act in any way they choose, it is essential to establish a governance framework that sets strict rules—which must be followed by the “higher-level” system—and ensures that decisions are made and actions are executed within an appropriate sandbox.
That is precisely why companies should not be focusing solely on SEO-driven presentation techniques designed to include their company name in AI-generated responses.
To use technology to secure a “proprietary space” that is not dependent on any specific platform and maintains a direct relationship between the company and its consumers. And to create a system that safely accumulates diverse and authentic VOC within that space and feeds it to AI in a controllable manner.
Rather than focusing solely on the algorithmic competition to be selected by AI, we will build up, safeguard, and pass on our own rich collection of “real human voices” so that AI can make the right choices for each and every individual.
No matter how much the times change, our mission remains the same: to listen to each and every person and continue creating a space where we can realize our ideal vision of “carpet sales.” I have a feeling that the arrival of the AI era presents us with a major opportunity to prove that ideal.
