This article was based on the interview with Taelor CEO Anya Cheng on AI, fashion, and sustainability by Greg Kihlström, Marketing and AI keynote speaker for The Agile Brand with Greg Kihlström podcast. Listen to the original episode here:
For those of us in the marketing world, we live with a fundamental tension. On one hand, the north star of modern MarTech is hyper-personalization, delivering the perfect message, product, and experience to an individual at the exact right moment. On the other hand, this relentless drive to cater to every whim often fuels a culture of hyper-consumption, a cycle of buying, using, and discarding that is increasingly at odds with a world waking up to its environmental limits. The fashion industry, with its seasonal cycles and the specter of “fast fashion,” is perhaps the most visible arena for this conflict. We are told to express our unique identity through what we wear, and yet the system that enables this expression contributes 20% of the world’s polluted water. It’s a paradox that many brands and marketing leaders are struggling to navigate.
The prevailing wisdom suggests you can either optimize for consumption or optimize for sustainability, but rarely both. However, new business models, supercharged by AI and a more sophisticated understanding of data, are starting to challenge this binary choice. What if the technology we use for personalization could also be the key to unlocking a more sustainable and efficient system? This is the very premise behind Taelor, an AI-driven men’s clothing rental and styling service. I recently spoke with its CEO, Anya Cheng, whose background at behemoths like Meta, eBay, and Target gives her a unique perspective on applying enterprise-level logic to a startup with a purpose-driven mission. Her approach demonstrates how a brand can build a powerful moat not just with a clever algorithm, but with a business model that generates a unique and invaluable dataset—one that serves the customer, the business, and the industry at large.
The New Moat: From Algorithms to Proprietary Data
For the past decade, the tech conversation has been dominated by algorithms. The black boxes at Google, Netflix, and Amazon that magically surface the right product or show have been the holy grail. But as large language models (LLMs) become more commoditized, the competitive advantage is shifting. The new frontier isn’t just about having a better algorithm; it’s about feeding that algorithm with better, more unique data that no one else can replicate. Cheng argues that we are at a critical inflection point, moving from an era defined by processing power to one defined by proprietary information.
“The algorithm era has ended. If you look at last, uh, era was all about algorithm… But the next 20 years in the era of AI were all mostly about unique data. If you don’t have proprietary data, then you’re just easy to replace by any large language model out there.”
This is a crucial insight for any marketing leader building their data strategy. A simple purchase history is no longer enough. Taelor’s rental model is, at its core, a sophisticated data-gathering engine. It captures information that a traditional retailer never could. When a customer buys a shirt at a 50% discount, does the brand know if they truly love the shirt, or just the price? Taelor’s model, much like Netflix’s subscription, decouples preference from price. More importantly, it creates a feedback loop on the physical product itself. As Cheng explains, “After wearing five times, this garment shrink. After washing five times, this garment become more loose.” This is invaluable, real-world quality assurance data captured at scale—information that is nearly impossible for a brand to obtain post-purchase. This blend of explicit preference data and implicit quality data creates a proprietary dataset that becomes the company’s true defensible asset.
Selling Outcomes, Not Products
Another core tenet of modern marketing is the shift from selling products to selling outcomes. We’ve all heard the cliché about selling the “hole,” not the “drill.” Yet, it’s a principle that is often difficult to execute at scale. How do you personalize for a customer’s ultimate goal, not just their demonstrated preferences? Taelor tackles this by combining AI-powered selection with a human-in-the-loop approach, focusing its intake process not on colors and patterns, but on goals. The customer isn’t just looking for a shirt; they are looking to close a deal, impress on a date, or fit in at a new client’s office.
“The goal for us is not just dress you up. It’s dress for people to achieve the final goal. Just like in marketers, we all say, like, we are not buying Coke, we are buying the freedoms. We are not buying Tesla… we are buying the status of that I’m a techie and I’m cool and I’m green, right?”
This is where the combination of AI and human stylists becomes so powerful. An AI can parse 30,000 garments to find items that match a user’s size, stated preferences, and the context of their goal (“conference,” “date night”). But a human stylist can make the intuitive leap to send a salesperson a shirt with a subtle dolphin print, knowing it will serve as an icebreaker to help him achieve his ultimate goal: closing a deal. This “dress for the goal” strategy has profound implications for how we measure success. Instead of simply tracking conversion rates on a product page, Taelor’s success is tied to the customer’s success and, ultimately, their retention. By understanding the “why” behind the “what,” they can build a service that feels less like a transaction and more like a partnership, significantly increasing customer lifetime value.
Solving the Upstream Problem: From Recycling to Prediction
The conversation around sustainability in retail often centers on the end of a product’s life: recycling, reselling, or renting. While these are important, they address the symptom, not the cause. The much larger, and more impactful, problem lies upstream in the supply chain: production. Cheng highlighted a staggering statistic that should give any leader in a product-based business pause. The industry is fundamentally broken at the point of creation, leading to immense waste before a product ever has a chance to be sold.
“Today 40% of clothes in the world goes unsold. Four zero, and 30% of them goes to landfill. So if the fashion brand continue to produce junk, the world is not going to be better. But why they produce junk? They don’t mean to, because they are designing something… two years down the road. And when they design, the only information… they use, was last year, what are things that sold?”
This is where Taelor’s B2B and B2C models converge brilliantly. The proprietary data gathered from their consumer rental service—what people with specific goals actually like to wear, and how those garments hold up over time—becomes a powerful predictive tool for brands. Instead of relying on flawed historical sales data, which Cheng notes is often a “really bad prediction of what’s going to sell next year,” brands can tap into a real-time stream of in-market feedback. This feedback loop has the potential to fundamentally change how clothing is designed and produced, moving the industry from a “make and sell” model to a “sense and respond” one. For marketing leaders, this is the ultimate realization of a data-driven strategy: using customer insights not just to optimize a campaign, but to inform the very creation of the product itself, reducing waste and increasing profitability in the process.
The Taelor model offers a compelling blueprint for the future, not just for fashion but for any industry grappling with the dual pressures of personalization and sustainability. It demonstrates that the path forward lies in building business models that treat the customer experience as a data-generation opportunity. By shifting the focus from a singular transaction—the sale—to an ongoing relationship, companies can unlock a richer, more contextual understanding of their customers and the performance of their products in the real world. This deeper understanding is the fuel for a more intelligent, efficient, and ultimately more sustainable ecosystem.
As leaders, we must challenge ourselves to think beyond optimizing the systems we have. We should ask: what data are we failing to capture in our current customer journey? Is there a service or a model that could not only provide more value to our customers but also generate a proprietary data asset that solves a larger, upstream problem in our industry? The models of tomorrow are being built today, not on algorithms alone, but on a deeper, data-informed understanding of human goals and product realities. It’s a complex and exciting challenge, and one that puts marketing squarely at the center of both business strategy and purposeful innovation.
