Milena Traikovich is a distinguished leader in demand generation and digital transformation, known for her ability to turn complex data into high-performing retail strategies. With a background rooted in analytics and performance optimization, she has spent years helping major brands navigate the shift from traditional merchandising to tech-driven lead generation. In this discussion, she explores how predictive tools are reshaping the fashion industry, using the “AI Trend Brain” as a blueprint for the future of apparel. We delve into the mechanics of trend identification, the reduction of inventory risk, and the evolving role of the creative professional in an era dominated by data-driven decision-making.
Retailers are now using social media and runway imagery to predict fashion demand. How do you integrate these diverse data points into a cohesive design strategy, and what specific metrics help you distinguish a fleeting viral moment from a sustainable seasonal trend?
Integrating these diverse data points requires a robust system that can synthesize social media content, runway images, and historical purchasing trends simultaneously. We look for patterns where a specific silhouette or print appears across multiple high-influence platforms, signaling that a trend is gaining genuine momentum. To distinguish a sustainable trend from a viral flash, we monitor engagement longevity and cross-category penetration rather than just a sudden spike in mentions. When a design concept shows consistent growth in digital interest and aligns with broader lifestyle shifts, it moves from a “maybe” to a core part of our design strategy. This approach allows us to act on data with confidence, ensuring our owned brands like Wild Fable remain relevant in a volatile market.
In categories like swimwear, silhouettes and print patterns can shift rapidly. How do you accelerate the development cycle from several months to just a few weeks, and what steps are necessary to ensure that quality remains consistent during such a high-speed production run?
Accelerating the development cycle involves adopting a digital-first model that streamlines the path from trend identification to product availability. By utilizing smaller production runs and direct-to-consumer shipping, we can bypass the lengthy timelines associated with traditional bulk manufacturing. This agility is especially crucial in swimwear, where we must capture seasonal demand before the window closes. To maintain quality, we rely on standardized production protocols and trusted vendor partnerships that can handle rapid turnarounds without compromising the integrity of the garment. This high-speed framework allows us to react to real-time feedback while maintaining the high standards our customers expect from our signature brands.
Forecasting demand for owned brands carries significant financial risk regarding excess inventory and markdowns. When a system flags a specific design early, how do you adjust purchasing decisions across different regions to minimize exposure to underperforming styles while scaling up high-performers?
When our AI flags a specific design, such as a particular polka dot pattern, we immediately pivot our buying strategy to lean into that success while it’s still early in the cycle. This allows us to scale up high-performers and secure inventory where demand is highest, effectively “buying” into the trend before it peaks. Conversely, we use these insights to identify styles that aren’t gaining traction, allowing us to reduce our exposure and minimize the risk of costly markdowns later. Effective inventory management is critical, especially considering that net sales recently fell 1.7% to $104.8 billion, making every purchasing decision vital to the bottom line. By being surgical with our inventory levels, we mitigate the business risks associated with inaccurate demand predictions.
Testing designs through direct-to-consumer shipping allows for real-world validation before a full retail rollout. What criteria do you use to move a product from a small digital test to wide-scale store distribution, and how does this digital-first approach change the role of the traditional designer?
We move a product from a digital test to wide-scale distribution based on clear performance indicators like click-through rates, conversion, and return frequency from our DTC channels. If a product resonates strongly in the digital space, it serves as a green light for a broader rollout across our physical store network. This digital-first approach doesn’t eliminate the designer’s intuition; rather, it provides them with a “live” laboratory to test their creative visions. Designers still spend significant time researching trends, but they now have a data-driven validation loop that informs their choices. It transforms the designer into a strategic partner who balances artistic flair with the hard reality of consumer behavior data.
With digital sales showing growth even as store-originated sales fluctuate, technology is becoming a central pillar of merchandising. How does a digital-first forecasting model improve the overall customer shopping experience, and what technical hurdles must be cleared to implement these systems across every apparel category?
A digital-first forecasting model improves the shopping experience by ensuring that the right products are in stock at the right time, reducing the frustration of “out of stock” messages. Since digitally originated comparable sales rose 1.9% while store-originated sales fell 3.9%, it is clear that our customers are increasingly looking for a tech-enhanced journey. The primary technical hurdle is integrating disparate data sources from across the entire company into a single, cohesive AI engine that works for everything from denim to basics. We must move away from the “standard” product development process for all categories and build a unified infrastructure that can process high volumes of real-time data. Overcoming these hurdles will allow us to personalize the experience even further, making our merchandising more responsive to individual shopper needs.
What is your forecast for AI-driven fashion forecasting?
My forecast is that AI will transition from a specialized testing tool to the foundational operating system for the entire fashion industry. We are moving toward a future where predictive models will virtually eliminate the guesswork in merchandising, allowing retailers to operate with near-perfect inventory levels. This will not only drive profitability but also lead to a more sustainable industry by significantly reducing the waste associated with unsold apparel. As these tools become more sophisticated, we will see a hyper-localized approach where store assortments are automatically adjusted based on neighborhood-specific data. Ultimately, AI will empower brands to be more creative and more responsive, closing the gap between a designer’s vision and the consumer’s wardrobe.
