1. WGSN: From Expert-Driven Forecasts to AI Integration

WGSN (Worth Global Style Network) is emblematic of the industry’s evolution. In the late 1990s and 2000s, WGSN built its reputation on traditional forecasting methods, albeit delivered in a then-new digital format. Its team of trend experts and researchers would travel the globe, attend fashion weeks, visit trade fairs, and analyze cultural trends, compiling their findings into reports for subscribers. WGSN’s early offerings included forecasts on apparel, accessories, consumer lifestyle trends, and even sketches and patterns designers could incorporate​. This model was largely expert driven: WGSN acted as a mega trend book, curating insights from human observation. For example, WGSN’s popular seasonal color forecasts were based on its analysts’ interpretation of societal moods and runway color palettes (e.g. predicting a rise of “Neo Mint” or “Cassis” two years out) – essentially a qualitative art backed by some research and historical knowledge) (WGSN official website). 

By the mid-2010s, WGSN began augmenting its process with data analytics. A merger with a competitor (Stylesight) in 2013 brought in new technology to its platform​. The real turning point came around 2019–2020, when WGSN invested heavily in machine learning, launching its AI-powered Trend Curve tool​. Mike Burgess, WGSN’s VP of Product, described this as bringing trend prediction “into the AI age”​​. WGSN did not abandon its human forecasters – rather, it combined their expertise with big data. The company assembled a “data nerd’s dream” of inputs: social media posts from 2,000+ influencers, a proprietary database of 22 million fashion images tagged by experts, detailed retail sell-through data, and even an image recognition system to detect precise colors (using Coloro codes) in pictures (Kershaw, 2021​). 

With Trend Curve, WGSN’s approach flipped to “machine-first, human-validated.” Instead of solely human hunches being later supported by data, now algorithms scan the data for interesting signals and humans verify or contextualize those findings​. For instance, rather than an analyst hypothesizing a trend and then checking data, the AI might flag an odd uptick in, say, pastel bucket hats across influencer posts, and then analysts would investigate its significance.

WGSN’s integration of AI has improved the granularity and actionability of its forecasts. As reported, WGSN can now forecast not just broad trends but specific items and details – “what details and trims are going to be big in womenswear in S/S 2021,” as Burgess noted​. The AI can quantify trend trajectories (when a trend will peak or fade) and even estimate the “volume” of a trend (how big it will get, and which consumer segment will drive it)​. Clients use these insights to make concrete decisions: how much to buy into a trend, when to stock it, and when to exit. An example outcome of WGSN’s AI era was validating the longevity of certain key colors from a seasonal palette – confirming that hues like Neo Mint and Purist Blue would have enduring appeal beyond one season, whereas a color like Mellow Yellow was peaking and should be used more cautiously. This gave brands confidence in which colors to emphasize or downplay in their upcoming lines (Sourcing Journal, 2024).

Importantly, WGSN still champions a human-AI partnership. The company states it “uses expert trend forecasting combined with data science”​– an approach echoed by its competitors as well. Francesca Muston of WGSN emphasizes that humans bring a wider context and creative interpretation to make the AI findings truly useful​. In practice, WGSN’s team of futurists and strategists (its Mindset consultancy arm) work with clients to translate Trend Curve data into brand-specific strategy. The WGSN case study highlights how a traditional forecaster can successfully evolve, leveraging AI to increase speed and accuracy while mitigating the AI’s lack of cultural nuance by keeping experts “in the loop” (Business of Fashion, 2024).

2. Heuritech vs. Traditional Forecasting Teams

Heuritech offers a contrasting case – a company born in the AI era (founded in 2013) that approaches trend forecasting with a tech-first mindset​. Heuritech’s premise is to “bridge the gap between AI and creativity” in fashion forecasting​. Its platform uses proprietary image recognition to analyze huge volumes of user-generated content. Rather than employing a large team of trend spotters, Heuritech’s core is its algorithm and data pipeline. It scans ~3 million fashion-related social images daily​, identifying products (shoes, dresses, accessories), colors, textures, and silhouettes. A key innovative aspect is how Heuritech segments the data by audience type. They categorize social media fashion data into three consumer groups: “edgy” early adopters, “trendy” mid-stage adopters, and “mainstream” followers​. By tracking which styles appear among edgy influencers versus the broader public, Heuritech can detect a trend in its incubation phase. Julie Pont, Heuritech’s fashion director, explains that slight changes in activity among edgy users are treated as early signals – the spark that might later catch fire in mainstream fashion​. This method is essentially a data-driven take on the classic “trickle-down” or “bubble-up” theories of fashion diffusion, quantifying how a look moves from avant-garde circles to mass appeal.

In a traditional forecasting team, say within a brand or at an old-school agency, you might have individuals assigned to monitor haute couture shows, street style, pop culture, etc., who then convene to identify common threads. Heuritech automates a large part of this monitoring and uses algorithms to connect the dots. The advantage is speed and evidence: Heuritech can quickly affirm whether an observed style has traction (by seeing its growth rate across thousands of posts and multiple cities). The disadvantage, as discussed, is that the human creative filter is lighter – although notably, Heuritech does employ human fashion experts to interpret the data and present it to clients (it’s not purely self-serve AI). Pont herself advocates for the “marriage of data and trend forecasting,” noting that data provides a strong backbone but “does not show you how to use it” – you still need creative strategy to apply the findings. In other words, Heuritech’s stance isn’t to replace human forecasters, but to supercharge them with better information.

A comparative scenario can be imagined: a traditional trend forecasting team at a fashion brand might consist of forecasters and designers who meet regularly, discuss observations (e.g., “lots of neon sneakers on the streets lately”), and consult external trend reports. If that brand uses Heuritech, they instead receive a data-rich report showing, for example, neon sneakers up 120% in edgy consumer posts in Korea and 50% in Europe, predicted to go mainstream in 6 months (hypothetically). The traditional team would rely on gut feel and maybe sales reports to decide if neon sneakers are worth investing in; the Heuritech-guided team gets a more quantified outlook. The result is often described as more “informed decision-making”​. Indeed, brands using Heuritech have concrete examples of aligning their decisions with data: Dior used Heuritech’s analysis to develop a new iteration of the classic Miss Dior handbag, essentially validating that certain design elements would resonate with current consumers. In another case, Paco Rabanne (Rabanne) consulted Heuritech to understand reactions to its reissued 1969 chainmail bag, helping the brand adjust marketing and production based on real-time feedback data.

So, is Heuritech “better” than a traditional team? It appears to excel in areas requiring breadth and quantification – scanning the global zeitgeist and giving numeric forecasts of trend evolution. A traditional team, however, might excel in deep creative synthesis – understanding why a trend fits the brand’s identity or how to put an inventive twist on it. Notably, Heuritech’s clients often includetraditional forecasting professionals at brands who use its reports as an input. Companies like Louis Vuitton, Prada, and New Balance have used Heuritech’s services to complement their internal teams​. New Balance, a sportswear brand, offers a perspective from the client side: Brad Lacey, NB’s design director of lifestyle footwear, noted that understanding people’s needs through AI data is “a big part” of making the business more sustainable and efficient. He suggests that traditionally “we don’t understand consumers’ desires… to the degree that we need to,” whereas data can fill those gaps​. This indicates that even for experienced teams, AI-based tools like Heuritech are valued for providing objective evidence and breadth that human forecasters alone might lack.

In summary, Heuritech vs. traditional teams is not a zero-sum contest; rather, it exemplifies how AI startups have pushed the industry to blend art and science. Traditional forecasters bring seasoned intuition and creative vision, while Heuritech brings computational power and scalability. Many in the industry conclude that the ideal formula is combining the two: as Pont remarks, “the combination of creativity and data [is] perfect” because the algorithm finds patterns, and the human decides what to do with them​. This collaborative stance is increasingly common in fashion companies today.

3. Other Examples from High Fashion, Sportswear, and Mass Market

The impact of AI in trend forecasting can be seen across various tiers of the fashion industry, from luxury houses to fast-fashion retailers and sportswear brands. Below are a few illustrative examples and case studies:

High Fashion (Luxury Brands): Luxury fashion was initially thought to be all about creative directors’ vision, possibly resistant to “data.” Yet, even high-end brands have embraced trend analytics to stay ahead. For instance, Dior and Louis Vuitton – storied Parisian maisons – have tapped into AI trend forecasting. Dior’s use of Heuritech data for the Miss Dior bag design has been noted, demonstrating that even a heritage brand values insight into what younger, digitally-engaged consumers want​. Louis Vuitton, along with Prada, also reportedly collaborates with Heuritech to shape upcoming projects​. This suggests that these luxury brands supplement their artistic process with data on emerging patterns (such as which styles of handbags or sneakers are gaining traction globally). The goal is not to dilute creativity, but to de-risk product decisions – e.g., choosing which archival design to reissue based on evidence of renewed interest (The Guardian, 2023). Another example is Marni and Miu Miu (Italian luxury labels): they have used data analytics platforms to identify opportunities in their collections​​. For Miu Miu, an accessories designer used a tool called NextAtlas to analyze social media behavior around themes like gender-fluid fashion and veganism, helping her see trend directions beyond her personal perspective​. These cases show that high fashion is increasingly data-informed. However, luxury brands still balance this with intuition – they often use data to validate a hunch rather than to generate the hunch. As Shinola’s CEO put it, “You never design by data, but the data provides a compass as you’re navigating a hunch.”​. In high fashion, the creative vision remains paramount, with AI as a supportive tool to ensure the vision aligns with (or deliberately challenges) the emerging consumer context (McDowell, 2019).

Sportswear and Streetwear Brands: Sportswear companies, which often sit at the intersection of performance and fashion, have been early adopters of data-driven approaches. New Balance we discussed, seeing AI forecasting as a way to better meet consumer needs and reduce wasted product​. Adidas has likewise been linked to AI trend forecasting – Heuritech counts Adidas as a client, and Adidas has experimented with AI to analyze fashion influencers (Adidas, Dior, and others were mentioned as early customers of Heuritech’s technology)​ (James, 2024). By analyzing streetwear trends on social media, Adidas can get feedback on styles like sneakers or athletic apparel silhouettes that are resonating with trendsetters. Nike, the U.S. sportswear giant, built an internal data science team and acquired startups to strengthen its trend and demand forecasting. Nike employs data scientists and AI models not just for inventory but to glean trend insights from its vast digital community (for example, tracking which sneaker designs gartner hype online vs. which fall flat). In fact, Nike, along with H&M and Dior, has been cited as a fashion leader investing in AI to personalize and predict trends (The Impression, 2025​). Sportswear brands also often leverage collaborations and cultural moments – consider Nike’s collaborations with off-white or Adidas with Kanye West’s Yeezy line. These can create trends in themselves. AI trend analysis helps sportswear brands identify what collaborations or retro styles might succeed by analyzing chatter and sentiment in the sneakerhead and streetwear communities. One interesting analysis looked at the convergence of luxury and sportswear: collaborations like Nike x Louis Vuitton or Adidas x Prada. Using AI, Heuritech studied these collaborations to understand their roots and reception​ (Heuritech, 2024). The data-driven insight can guide brands on how to blend high fashion aesthetics with athletic wear in a way that authentically connects with youth culture. Overall, in sportswear, trend forecasting with AI is used to ensure that design teams are aligned with fast-moving street trends (like the rise of athleisure or the resurgence of 90s sports fashion) and to know early if a particular style (chunky “dad” sneakers, for example) is accelerating or fading worldwide.

Fast Fashion and Mass Market Retailers: Perhaps nowhere is trend forecasting more critical than in fast fashion and mass retail, where getting the trend right can make or break sales for the season. Companies like Zara and H&M historically used a combination of catwalk monitoring and quick-response supply chains to ride trends (Thomasnet, 2020). Today, they heavily incorporate AI. The fast-fashion model is to identify a trend and have it in stores in a matter of weeks; AI makes identifying those trends almost instantaneous. H&M Group, for example, employs over 200 data scientists and advanced AI algorithms specifically to predict and analyze trends. By crunching social media trends, store transactions, and even loyalty program data, H&M’s AI team can foresee demand shifts and guide designers on what styles to produce. The former CEO of H&M noted that their AI initiatives had “already helped the company predict trends” by analyzing customer behavior alongside societal shifts​. 

This has a sustainability angle too: better trend prediction helps avoid overproduction (making what people will actually buy, and less of what they won’t)​ (H&M Group, 2023). Zara, on the other hand, leverages its near-real-time sales data across its thousands of stores and online to gauge trend performance. While Zara is famously secretive, reports indicate it uses AI to sort customer feedback and social media signals to inform its designers which cuts or colors are gaining momentum​​ (Inditex (Zara), 2023). Fast fashion retailer Shein (China) takes this to another level: by obsessively mining social media hashtags, influencer posts, and search trends, Shein can identify a micro-trend (say a certain dress style popular on TikTok) and produce it almost immediately​. Shein’s use of data and AI is so aggressive that it often can release new styles faster than competitors, essentially riding the social media algorithm waves in real time (Shein, 2024). This ability to turn a “blink-and-you-miss-it” meme into a product is a direct outcome of AI-powered trend tracking.

Multi-Channel Retailers and Department Stores: It’s not just the producers of fashion, but also retailers and buyers who use these forecasting methods. Large retailers like Macy’s or Target in the U.S., or Marks & Spencer in the UK, have buying teams that historically relied on trend reports (from agencies like WGSN or Trend Union) and their own experience. Now, many of them use data platforms. For example, the department store Debenhams’ forecaster Angela Baidoo uses EDITED, a retail analytics tool that scans online assortments worldwide, to see what products and styles competitors are stocking and selling in real time (EDITED, 2023)​. This gives buyers a competitive edge – an immediate picture if, say, puffer jackets in a certain color are selling out everywhere, indicating a hot trend. Similarly, e-commerce companies like ASOS rely on AI to tailor their offerings to trends as they emerge (ASOS being digitally native, it tracks search and browsing data closely to inform trend picks). In these cases, the trend forecasting function is deeply integrated with demand forecasting and inventory – a trend isn’t just a concept but tied to how many units of a style to order (McDowell, 2019), (ASOS, 2024).

These examples underline a few key takeaways. First, AI-driven trend forecasting is pervasive – from haute couture ateliers to bargain fast-fashion chains, everyone is tuning in to data for trend signals. Second, the marriage of traditional and AI methods is often the norm: brands subscribe to agencies like WGSN or Fashion Snoops for a broad creative outlook, and use AI analytics tools for granular, immediate insights (McKinsey & Company, 2024). A design director candidly shared that her team uses in-house experts plus multiple external resources: traditional forecasters (Peclers, Nelly Rodi, Trend Union) and a data analytics service (StyleSage) coupled with customer sales info​ (Retail Dive, 2023). This blend ensures they cover all bases – the creative, the cultural, and the commercial data. Third, we see that sportswear and streetwear brands place emphasis on early detection of trends bubbling up from youth culture, which AI is particularly adept at capturing, whereas luxury brands use AI to refine and validate rather than originate trends. Mass-market players, driven by profit margins, use AI to minimize guesswork and maximize hit rates on trend-right products (NellyRodi, 2023), (Business of Fashion, 2024).

Over the past several decades, fashion trend forecasting has shifted from an almost mystical art practiced by industry “oracles” to a tech-empowered science guided by big data – yet, it remains in many ways an art-and-science hybrid. Traditional methods gave us the foundational understanding of how trends form, spread, and cycle, relying on human creativity, deep cultural knowledge, and intuition. These methods excel at providing narrative and meaning – understanding not just what might be popular, but why it will resonate. Their drawbacks — slow speed, limited scope, and potential bias — have been increasingly addressed by the advent of AI-powered forecasting. AI brings speed, scale, and analytical precision, turning the whole world (and web) into the forecasting playground. Real-time social media scraping, search trend analysis, and retail data mining have made trend detection more immediate and, in many cases, more accurate in predicting commercial success (Business of Fashion, 2024). 

However, the literature and case studies make it clear that AI has not rendered human forecasters obsolete. Instead, the role of the human has evolved. As WGSN’s experts and Heuritech’s team both affirm, the synergy of human + AI is the optimal path. AI can rapidly map out the landscape of “what is trending now,” identify patterns and even project them forward, but it takes human insight to connect those dots to larger cultural movements, to filter out the noise, and to envision which trends align with a brand’s identity or could captivate consumers’ imagination in the future. We see this synergy in practice: a machine might tell WGSN that cardigans are spiking in data, but it took human curiosity (noticing Harry Styles’ cardigan moment and a viral TikTok knitting trend) to ask the machine to look at cardigans in the first place​. Conversely, a human forecaster might feel that “escapism will be a big post-pandemic theme,” and then use AI to validate that by checking if travel-inspired prints or bold colors are indeed gaining traction in online data (WGSN, 2024).

In comparing traditional vs. AI methods across speed, scale, predictive power, and adaptability, it’s evident that each compensates for the other’s weaknesses. Traditional forecasting’s slower pace and subjectivity are countered by AI’s fast, expansive analysis​. AI’s lack of nuance and context is countered by human interpretation and creative vision. The evolution from the pre-2000s era to today has not been a replacement of old with new, but rather an integration. WGSN’s journey from trend books to Trend Curve AI, and the widespread use of hybrid strategies at brands, exemplify this blended approach (Lu, S., & Kershaw, M. (2024).

Looking ahead, sources suggest this trend of integration will only deepen. As one report notes, generative AI and advanced analytics could add over $150–$275 billion to the fashion industry in the next 3–5 years​, indicating massive investment and trust in these tools. We can expect even more micro-trends and personalized trend insights, as AI parses niche communities and even individual consumer preferences. This could lead to a future where brands create hyper-local or subculture-specific collections informed by data. On the flip side, concerns about sustainability and creativity remain: will AI-fueled micro-trends accelerate consumption, or will they enable smarter planning and less waste? The answer likely lies in how responsibly companies use these tools. If they use AI simply to chase every trend, we may see burnout and backlash; if they use AI to better plan and to inform creative innovation (not stifle it), the outcome could be positive for both business and creativity (McKinsey & Company, 2025).

In conclusion, traditional and AI-powered forecasting are not mutually exclusive but complementary. The fashion industry is learning to leverage the best of both: the responsiveness and scale of AI with the refinement and foresight of human experts. The result is a more agile, informed forecasting process than ever before, one that is still grounded in the core goal that has always driven trend forecasting – helping brands anticipate the future needs and desires of consumers. And as the case studies from couture to sportswear show, those who master this new blend of art and science in trend forecasting are positioned to thrive in fashion’s ever-changing landscape.