Artificial Intelligence (AI) has become a buzzword across industries, including fashion. Yet, while many companies are excited about the potential of generative AI and predictive tools, few are realizing significant returns on their investments. According to a recent report from Boston Consulting Group, only 4 percent of CEOs polled felt their investments in generative AI delivered measurable value. That’s a paltry sum given the myriad ways AI could help solve many of the industry’s challenges—from sizing and inventory allocation to logistics and capacity utilization.
Here, Dr. Ahmed Zaidi, co-founder and CEO of AI platform, Hyran Technologies, shared his insights on what’s new in AI technology, how to close the gap between potential and implementation, and how to use AI to optimize supply chains for disruption.
HGI: How are apparel companies using AI right now, and what’s holding them back from making substantial progress?
AZ: There’s a lot of hype around AI — generative AI, predictive AI, and newer terms like agentic AI. While the excitement is palpable, it’s important to understand that AI itself is not a single technology but more of an evolving destination of innovation. The current state of AI is heavily driven by deep learning, which has evolved from earlier AI technologies like rule-based systems and traditional machine learning.
Deep learning enables systems to abstract the “essence” of objects through massive data sets, but this creates challenges. Brands in the apparel industry want to use AI to solve problems like forecasting demand and managing their inventory, but they often lack the right data or processes to implement it effectively. For instance, forecasting models have existed for decades, but no brand has raised its hand to say inventory issues have been solved simply because deep learning is now in the mix.
The root problem is that many of the industry’s existing processes and planning systems—such as materials planning, production, and sourcing—were designed for an era that never accounted for AI-based technologies. AI, in many ways, is like attaching a jet engine to a horse carriage. Without a redesign of the underlying infrastructure, that jet engine isn’t going to make the carriage move faster; it’s just going to scare the horses.
HGI: Let’s talk more about this idea of redesigning processes. You mentioned apparel brands have certain ingrained habits—how does this affect their ability to use AI effectively?
AZ: A great example is how brands approach sizing. Most companies don’t plan down to an individual size level; they use distribution curves that group sizes into general percentages within regions. For instance, they might plan 25 percent as small, 35 percent as medium, and so on. This was efficient when forecasting tools were less advanced, as planning for size-specific SKUs would have involved too much manual complexity.
However, AI allows us to make decisions at much higher dimensions. We can plan production down to the level of specific product features. But to take advantage of this, brands must rethink their decision-making systems to plan and gather demand data at this more granular level.
It’s the same across the supply chain, too. Designers, for example, often make creative decisions they think “look nice,” without visibility into how these choices impact production and logistics. A small design tweak, like changing a fabric from 120 GSM to 125 GSM, can ripple throughout the supply chain — altering production timelines, costs, and supply chain operations like lab dips or fabric runs. Without rethinking these fragmented processes, it’s hard for brands to effectively implement AI systems.
HGI: We’re living in uncertain times. Supply chains are being affected by everything from geopolitical tensions to shifting consumer demand. Can AI truly help brands make better decisions in uncertain markets?
AZ: Yes, absolutely — under the right conditions. The challenge here is that most brands are trying to use AI to address uncertainty in ways that are inherently flawed. They are designing systems that assume they know exactly what the future looks like.
Forecasting processes in fashion, for instance, are driven by traditional material resource planning (MRP), which was developed for scenarios with clear certainties. At its core, these systems create forecasts based on single-number predictions: “We need precisely X units of Y product.” When things inevitably shift, the entire chain suffers, whether through excess inventory, missed demand, or delayed deliveries.
Instead, we encourage brands to adopt systems designed to be “approximately right” rather than “precisely wrong.” For example, when planning materials, we advocate for incremental purchases over time rather than making one big bet. This allows brands to adjust to evolving data and make more informed decisions. AI tools can handle the complexity of these flowing materials and provide actionable recommendations in real time—something a human supply chain manager or legacy system simply cannot do.
HGI: AI implementation requires brands to trust the technology. How do you help brands overcome skepticism and integrate AI effectively?
AZ: Trust is a huge factor. Many brands form AI committees to evaluate the models we build for them. These committees go beyond algorithms; they scrutinize how introducing AI impacts their overall operations and decision-making processes.
A key part of building this trust is transparency. For instance, we show brands the counterfactual — what would happen if they stuck with their traditional decision-making versus using our AI tools. By quantifying the differences in delays, costs, or excess production, brands start seeing the tangible ROI.
“AI is a jet engine in an industry that operates on a horse-and-cart infrastructure. For brands to truly benefit, they need to rethink their processes”
—Dr. Ahmed Zaidi, Hyran Technologies
We also educate brands extensively, helping them understand what the models are—and, just as critically, what they are not. It’s not merely about providing predictions but also about offering scenario planning. How should they react to unexpected changes like a buyer suddenly demanding earlier delivery or government tariffs disrupting their supply chain? AI can offer insights and recommend buffers for such volatility, but brands need a robust understanding of how to interpret and act on those insights.
HGI: You just mentioned external influences like tariffs. How does your platform address factors outside of a brand’s control?
AZ: This is where optimization comes into play, a rarely discussed but vital subset of AI in fashion. Optimization allows brands to evaluate multiple scenarios in the face of external shocks, whether it’s a new tariff, shipping delays or changing cost structures.
For instance, AI can help brands assess trade-offs. If a delay occurs, it can suggest adjustments across the supply chain—like reallocating volumes between factories or changes to shipping routes—to minimize both cost and lead times. You can even prioritize what matters most to your business, whether that’s delivery speed or saving costs, and the model will create scenarios to align with those priorities.
Ultimately, optimization complements predictive models. When a disruption happens, optimization provides the path forward.
HGI: Explain how the Hyran Technologies platform works and its core value proposition for apparel brands.
AZ: At its simplest, our platform simplifies executing the complexity of an agile fashion supply chain. It has two core components: the brand side and the factory side.
On the brand side, we turn demand forecasts into agile, actionable materials plans that minimize lead times, delays, and material liability. We start by breaking the forecast into subcomponents using the Bill of Materials (BOM). Our AI engine applies risk models to the forecast to determine the optimal quantity of material to position, based on the volatility of each SKU. The planning models incorporate lead times, minimum order quantities (MOQs), and the strategic priorities of each SKU to calculate when and at what stage (e.g. yarn, greige fabric, or dyed fabric) materials should be positioned. As forecasts are updated, the AI platform dynamically makes trade-offs and regenerates the material plan to align with the brand’s business objectives, such as on-time delivery and liability reduction.
On the factory side, we address the challenge of planning and replanning. Factories operate under constantly changing conditions—from shifting forecasts to external disruptions like worker absences, yarn delays, and quality issues. Our platform monitors these factors and dynamically replans the execution approach in real time. This helps factories optimise key objectives such as capacity utilisation, while maintaining agility and ensuring on-time delivery for the brand. We also relay this visibility back to brands, enabling greater alignment between both sides and helping to reduce delays, excess inventory, and missed orders.
The top-line value of our platform is agility and managing uncertainty. By helping brands and factories collaborate in real time, we enable them to be more responsive to market changes, which in turn reduces delays, increases efficiency, and minimizes costs.
HGI: Any closing thoughts you’d like to share?
AZ: I always emphasize this: AI is a jet engine in an industry that operates on a horse-and-cart infrastructure. For brands to truly benefit, they need to rethink their processes, prioritizing agility and collaboration over outdated systems of certainty and isolation.
Secondly, they need to embrace their supply chains as partners, not transactional entities. Factories, vendors, and other players in the value chain must be aligned to create a truly connected and optimized system.
Finally, brands need to think holistically about the cost of ownership, not just short-term margins. This means sometimes trading off efficiency (cost) with agility. By being willing to test, adapt, and engage AI strategically, they’ll not only be better equipped for uncertainty but also build healthier, more resilient businesses.
