Intermediaries or the so-called buying houses or liaison offices have been a stable feature of international trade in textiles and apparel since the nineteenth century. These institutional structures are based on mechanisms that are the antithesis of the collaborative buyer-supplier relationship, drawing in the popular distrust of ‘middlemen’, yet they are highly productive providing benefits to both buyers and suppliers.
Fundamentally, any intermediary organisation or buying house has five key roles to play which are Search, Specification, Negotiation, Completion and Enforcement. All these five roles involve information costs, only negotiation and enforcement involve transaction costs. It is pertinent to understand here that of the manifold functions that a buying house serves at any point of time, it is the information cost that forms its very existence because buyers and suppliers spend their valuable time, effort and money to avail the authentic information that can be solely supplied by buying houses. This cost forms the core of decision-making, problem solving and research carried out by the buying houses and is a part of the transaction cost that occurs during collecting, communicating and storing of information.
Let’s now delve into how a buying house functions based on the key roles defined above. Search function is the identification of right trend (colour/silhouette/material) to develop right merchandise that will sell and identification of right supplier/vendor for the right merchandise; this is pure information cost. Specification is techpack finalisation and negotiation involves cost and delivery date finalisation; again pure information cost. Negotiation and enforcement involve sampling execution and approvals, pilot production approvals, quality audit and warehousing; this is a mix of information and transaction cost. Interestingly all these functions are nicely re-worded by the leading organisations.
The world’s largest buying organisation Li & Fung defines Collaborate, Innovate, Source, Oversee and Deliver as its key functions. India’s most valuable buying organisation Triburg specifies Design services, Fabric R&D, Technical services, Quality control and Logistics as its key functions. Also the buying houses no longer identify themselves as intermediaries; for example, Li & Fung now calls itself a supply chain solutions company.
It must be noted here that of the five key roles mentioned above, search, specifications and negotiations occur before the order execution starts whilst completion and enforcement occur post- execution. Merchandising is at the epicentre of search-specifications-negotiations functions and nearly 40 per cent of the workforce is engaged in the job. But are these roles under threat from silent invasion of AI?
Let’s see how the search function is executed by a typical merchandiser in any buying house. The right merchandise at right time at the right cost may sound simple; but the actual parameters to be considered are numerous. The different vendor rating systems can consider as many as 50 parameters across different domains like product, business, financial, legal, digital and ESG (Environmental, Social and Governance). It is impossible for a human merchandiser to optimise decisions considering all these parameters and technology has to come in rescue here. Either BI (business intelligence) or AI (Artificial Intelligence) have to work here. BI has less probability of success as it considers only algorithmic functions to optimise (based on past data to make future decisions) and the human touch is missing.
Currently, such number crunching is done in the brain of highly experienced, highly accomplished, highly paid merchandisers and without an iota of any optimising mathematical function. Some of the progressive intermediaries in recent years have however attempted some algorithmic approach, but here comes the disruptive AI to the rescue. Can AI take the role of complex multi-priority optimisation decision-making? The answer is unfortunately yes. Does the recent lay off and/or hiring freeze in many liaison offices of brands and retailers across Asia indicate any ominous trends?
Unlike BI, the AI can consider the human experience and intuitions as well to take human-like decisions. Years of performance data from hundreds of fabric suppliers, thousands of apparel manufacturers and numerous attributes of each of them are fed to ML (machine learning) or DL (deep learning) systems, so that the AI is able to analyse and reach the elusive dream combination of right supplier to right buyer to achieve low cost yet high quality and timely delivery.
Most of us may recollect the scandal of the 2010s, where personal data belonging to millions of Facebook users was collected without their consent by British consulting firm Cambridge Analytica, predominantly to be used for political advertising. In the aftermath, CEO Mark Zuckerberg testified in front of US Congress and Cambridge Analytica filed for bankruptcy. But who was the whistle-blower, whose revelation to The Guardian prompted the scandal? Where is he now?
The whistle-blower Christopher Wylie is a data consultant and incidentally has PhD in predicting fashions trends from the University of the Arts London, and most importantly now works as a consultant with one of the world’s top five fashion brands. The same fashion brand incidentally has a global technology cell with 100+ data scientists working on automation of many logical functions that require years of experience backed by data. Be it vendor selection, vendor rating, merchandise planning, demand forecasting, vendor mapping with merchandise sourcing and many more. But, weren’t these the jobs of experienced merchandisers earlier? The chilling yet hard truth is yes it was their job, but will no longer be in future.
During a conference in 2018, I had predicted that any function that follows algorithmic or logical functions will be automated sooner or later. The apparel manufacturing industry during the last three decades has tilted more towards management functions and less towards technical skills and I foresee these management functions to be at the risk of automation. The so-called ‘human experience’ is now replaced by ‘data analytics’ and the ‘information cost’ is reduced to ‘zero’. The AI/ML-based applications can be built to automate search, specification and negotiations jobs with better accuracy and faster efficacy. The icing on the cake is these applications don’t require expensive salary and perks, no attrition and yet they continuously learn and improve themselves over use.
AI application in retail and customer-focused space has grown multi-fold and there exist many solutions. If we track the recent acquisition, investment and partnership trends of some of the big retailers worldwide, the writing on the wall is clear. Technology and automation are on everyone’s radar; supply chain optimisation, omnichannel engagement and last-mile delivery are common picks. The age-old buying houses or liaison offices may have renamed themselves as ‘supply chain’ companies but can they withstand the onslaught of AI? Let’s wait and watch in 2023!
Bharat Tex Special Report 2026 
