How AI is reshaping the role of trend agencies in fashion
The announcement was made this week and marks a tipping point for the fashion industry: the acquisition of the historic trend forecasting agency NellyRodi by the Ifop group (French Institute of Public Opinion). This alliance between a pillar of creative forecasting and a giant in data and sociology signals a new era. This transformation has become a matter of survival. The recent liquidation of Promostyl, another pioneering trend agency in the sector, serves as a stark reminder. Forecasting is no longer (just) about mood boards. It now lies at the intersection of data, consumer behaviour and brand strategy.
As generative artificial intelligence democratises access to global data, trend agencies face an existential paradox. Since everyone uses the same AI models to analyse the market, how can they continue to sell foresight? Faced with the threat of algorithmic homogenisation, trend forecasting agencies are undergoing a profound transformation.
Illusion of AI's omniscience and the trap of standardisation
The arrival of LLMs (GPT, DeepSeek or Claude) has transformed the research phase. AI can simultaneously scan and organise articles, academic research, social conversations and product launches on a superhuman scale. This computing power, however, comes with a major bias: it flattens creativity.
"The goal is not to let the 'loudest voice' win, as AI might by giving more weight to the most visible content, but to build a collective forward-looking vision," states Anne Etienne-Reboul, CEO of the Peclers agency. She highlights the fundamental difference between trend reporting (reporting on what already exists, which AI does very well) and trend forecasting (anticipating what will emerge).
AI generates what the agency Fashion Snoops calls "baseline" trends, built on common data. However, as the agency points out, the historic value of forecasting has always been in divergence. Worse still, the algorithm lacks commercial instinct. When asked about this, fashion expert Fériel Karoui illustrates it perfectly: "AI can help sort through the chaos of social media (...) but it misses contextual and sentimental parameters. We need to analyse the relevance of the figures and distinguish between a viral media success that floods the internet and a future commercial success." She cites the omnipresence of skirts in menswear shows: a powerful symbol of deconstruction on social networks, but a highly improbable commercial success for all market players.
New profession: from forecaster to anti-noise curator
The profusion generated by AI has paradoxically made the human filter indispensable. Yesterday's major unifying trends have shattered. "Trends are increasingly fragmented," observes Fériel Karoui, pointing to the emergence of internet micro-communities (the "-core" or "-maxxing" aesthetics).
In this landscape saturated with "content slop" and fake news, the agency's role is changing. It is no longer about prescribing fashion, but about knowing which information deserves attention. As the Fashion Snoops team summarises in an email: "AI can surface a thousand signals, but you need a person who understands culture, attends trade shows, watches the catwalks and lives the industry from the inside to know which three signals matter and why."
Clients no longer pay to know "what's next," but to navigate the chaos. Bespoke service is becoming the norm. Faced with paradoxical trends and tighter budgets, the role of the trend agency is to identify, from the multitude, the exact movement that will resonate with a brand's specific DNA and its target audience.
Yet, despite the digitalisation of insights, the legendary trend book – that heavy binder filled with fabric swatches and silhouettes to illustrate seasonal themes – is not dead. Faced with the mass generation of synthetic images, the CEO of Peclers is categorical: "No AI in our trend books!" The agency insists on absolute traceability with precise bibliographies of its sources, and an unapologetic return to pure craftsmanship to counter the machine. The director also notes that her books are still consulted today by style, design, marketing, retail and forecasting teams, "who use them as tools for inspiration and anticipation".
Algorithmic counter-attack: closed AI and bespoke solutions
Far from rejecting technology, agencies are fighting back by developing their own hybrid ecosystems. They no longer use AI to scour the open web, but to intelligently query their own closed, historical databases. Peclers has developed its own internal tool, while Fashion Snoops is deploying conversational tools like "Ask FS" and "MUSE".
The revolution lies in this "Living Intelligence". Rather than delivering a static report that is valid for 18 months, agencies now allow their clients to track the migration of a signal in real time (from wellness to fashion, then to retail design). Above all, they are democratising access to these insights internally. With a simple natural language prompt, a marketing junior or a finance director can query 25 years of forecasting archives, guaranteed to be free of algorithmic hallucinations.
"The questions our clients are asking are changing. It's no longer 'what's happening?', but 'what matters to my brand?' and 'how can we better stand out?'" confirms Fashion Snoops. The technological tool is used for absolute personalisation, allowing designers to interpret a trend with their own palettes, materials and silhouettes.
While artificial intelligence has made information overabundant, it has made critical judgement extremely rare. By combining the rigour of data with human sensibility, trend forecasting agencies are proving that the algorithm is just a compass; only the cultural analyst knows how to read the map.