TOKYO - How will AI transform marketing? The answer, according to leading marketer Kazuki Nishiguchi, lies not in marginal efficiency gains but in a dramatic restructuring of business itself, as AI agents move closer to consumers and potentially displace even dominant platforms such as Amazon.
In a recent episode of the TV Tokyo program “Nishiguchi-san, Is Marketing Really Necessary?”, TV Tokyo announcer Emika Furuhata sat down with Nishiguchi, former marketing director at Procter & Gamble and now president of Strategy Partners and Wisdom Evolution Company, to examine Amazon’s growth strategy through the lens of “customer dynamics.”
Nishiguchi, who has advised more than 300 companies and is known for advocating “N=1 analysis” — a method focused on deeply understanding a single customer — argues that sustainable growth depends on identifying and cultivating “good customers” who generate “good revenue.”
At the core of his framework is a five-tier customer pyramid: loyal customers, general customers, lapsed customers, aware-but-unpurchased customers, and those who are completely unaware of the brand. In many businesses, loyal and general customers account for only 10% to 20% of the total potential market, while a substantial share remains either lapsed or unaware.
“The mistake many companies make,” Nishiguchi explained, “is to give up too early, assuming the product life cycle is over, when in fact a large pool of unaware customers still exists.”
Importantly, these categories are not static. Customers move between tiers in real time. Loyal buyers can become even more committed — increasing purchase frequency and lifetime value — or they can lapse. Lapsed customers may return if circumstances change or if reminded of a positive experience. The key, Nishiguchi said, is to monitor both the composition and the movement within the pyramid.
Amazon’s Two-Decade Flywheel
Amazon, he noted, offers a textbook example of this dynamic in action. Beginning with books in the early 2000s, the company steadily expanded into CDs, DVDs and then into an ever-widening array of product categories.
By continuously broadening its assortment, Amazon encouraged existing book buyers to purchase additional items — increasing both transaction frequency and average order value. The introduction of Prime membership and free shipping further reduced the friction that might otherwise have driven customers away.
“As categories expanded, customers had fewer reasons to leave,” Nishiguchi said. “Even if they no longer needed books, they could buy shampoo or camping gear.”
This strategy required enormous upfront investment in logistics infrastructure and IT systems. During the early 2000s, particularly in the aftermath of the dot-com bubble, Amazon posted large losses as it built warehouses and distribution networks. At the time, many investors doubted the viability of such aggressive spending.
But founder Jeff Bezos was betting on lifetime value. If customers who had already registered payment methods could be encouraged to shift more of their total spending onto Amazon, frequency and spending per customer would surge, eventually pushing the company past its break-even point.
History proved him right.
“The story was so grand that many people couldn’t see it at the time,” Nishiguchi said. “But the logic was clear: expand categories, raise frequency, and long-term profitability will follow.”
He contrasted Amazon’s approach with failed dot-com ventures such as Pets.com, which focused narrowly on selling pet products online. Without a structural mechanism to significantly increase purchase frequency or create strong differentiation, such businesses struggled to achieve sustainable profits.
The lesson for companies today, Nishiguchi said, is not to replicate Amazon’s scale, but to design offerings around long-term customer frequency rather than simply launching new products to chase short-term revenue.
“Think in five-year terms,” he said. “What services or products will increase how often customers engage with you? Frequency matters even more than price.”
The AI Agent Shock
Yet even Amazon’s powerful model may face disruption.
As generative AI advances, so-called AI agents — systems embedded in platforms such as Google’s Gemini and OpenAI’s ChatGPT — are beginning to mediate interactions between consumers and products.
Already, OpenAI has partnered with Shopify merchants to enable purchases directly through ChatGPT. Rather than visiting a retailer’s website, consumers can search, receive recommendations and complete transactions within the AI interface.
If such agents become the primary “front door” for consumers, platforms like Amazon could be relegated to back-end infrastructure — handling logistics and fulfillment while losing direct access to customer data and engagement.
“When an AI agent stays closely attached to the customer, there may be no need to visit Amazon’s storefront at all,” Nishiguchi said. “The agent can search, compare and execute the purchase.”
From the consumer’s perspective, the ideal outcome is simple: find the best product at the best price with minimal effort. Whether the transaction flows through Google, OpenAI or Amazon becomes secondary.
Google, with its dominant search traffic, possesses the technical capability to integrate commerce deeply into AI-assisted search. If it — or any major AI provider — systematically links user intent to fulfillment partners, the distance between desire and transaction could shrink dramatically.
In such a scenario, the competitive battlefield shifts. The company controlling the customer interface — and thus the initial data and intent signals — gains strategic leverage. Retailers and marketplaces risk becoming commoditized service providers.
For Nishiguchi, the development reinforces a broader point: business success depends on understanding structural shifts in customer behavior. Amazon mastered the art of expanding frequency and lifetime value in a pre-AI era. The next decade may belong to firms that understand how AI agents redefine proximity to the customer.
“Good revenue,” he said, “is revenue that becomes stronger over time. But if the point of contact moves elsewhere, even great companies must rethink their model.”
As AI agents draw ever closer to consumers, the question is no longer whether marketing is necessary — but how fundamentally it must evolve.
Source: テレ東BIZ
