AI in fashion: inventory optimization for fashion retail at MS Mode

interview
Artificial intelligence (AI) is changing the fashion industry at a rapid pace. While some companies are taking a wait-and-see approach, others are taking bold steps forward.

In this third part of the AI ​​in Fashion series, Dennis Mok, Chief Executive Officer (CEO) of Dutch fashion company MS Mode, explains how the company uses artificial intelligence to optimize inventory. He also explains what opportunities he sees for the future of the company and the fashion industry. MS Mode operates over 200 stores in the Netherlands, France, Belgium, Spain and Luxembourg.

How does MS Mode feel about AI and what is the company currently using it for?

Artificial intelligence, like any other innovation, is something that needs to be addressed. Our approach is very practical: what are the possibilities? And: Can this be turned into a business case? If so, it is wise to do it. After all, we are a fashion company: we make and sell fashion. For us, AI is not a goal, but a means.

MS Mode uses AI at various levels within the company. However, the most important application is in inventory management and demand forecasting. McKinsey’s 2018 State of Fashion report said demand forecasting is one of the biggest challenges for fashion companies, and that’s true. This is exactly the challenge Thunderstock focuses on. We were there early. The collaboration with the sister company America Today was already underway when I started as CEO at MS Mode at the end of 2019.

With the inventory module from Thunderstock, a Dutch software company, we can allocate inventory, i.e. distribute, replenish and relocate. This way we avoid surpluses and bottlenecks. This happens at the Stock Keeping Unit (SKU) level: per product variant, color and size.

Inventory is one of a fashion company’s largest investments. If you can make this more intelligent, you can make a huge profit.

What profits have already been achieved?

The benefits of demand forecasting are very clear. We can achieve the same sales with less inventory, which has a direct positive impact on the margin. Since we have to offer fewer discounts, profitability increases. This sets off an upward spiral: with less inventory you achieve better results, and these results in turn help you manage more intelligently. Of course there is a limit because you cannot optimize infinitely. But with AI you can get much closer to this optimum.

Inventory optimization is important not only for efficiency and results, but also for customers. Because they often find something that they really want. MS Mode operates in various countries, with stores in large cities to smaller shopping centers. We are also present online via our webshop and online platforms such as Zalando and Wehkamp. All of these customers have slightly different preferences and purchasing behavior. Thanks to AI, the offer better matches what customers are looking for. And that is perhaps the biggest advantage.

What else are you doing with AI?

We also use AI internally. All employees have access to ChatGPT Enterprise and are encouraged to actively use it. As with many innovations, the technology penetrates into everyday working life when people use it privately. Our very active data team helps us with this and provides personal support to the managers. They provide explanations and support in creating bots and agents to make processes smarter and faster. This ranges from checking or drawing up contracts to assessing the sales potential of new store locations, for example in France. We used to do the latter mainly based on feeling and experience, but today we combine that with data.

Our customer service is working with mobile service provider CM.com on an AI customer service agent that will answer the most frequently asked questions from customers. This is currently being tested and will go live in the webshops shortly. In addition to chat, telephone and email, we can help customers even faster and more specifically.

We are also exploring the use of AI in our business intelligence environment. We want to further automate reports and analyzes and enrich them with AI insights. This means we can control more quickly based on current data.

We also use machine learning in our webshop, for example for product recommendations, personalized sorting and automatic messages. In this way, we can address our customers in a more relevant and personal way.

What have you learned so far from using AI? And what does that require of your employees?

Thunderstock was one of the first systems on the market with this technology. Because we applied it directly internally, we learned a lot about how to use it effectively. In the beginning you have to figure out how to make it more user-friendly. The reports also need to be better adapted to practice. However, the biggest challenge was not in the technology, but in the organization itself. The same applies to AI as to everything new: new things always meet with resistance.

Like every fashion company, we also have our fixed ways of working: planning, purchasing, distributing. And then suddenly a system comes along that helps and sometimes also takes over tasks. This requires adjustment and trust. You have to learn to trust that the system’s predictions are correct. Sometimes people think, ‘We’ve always done it this way, and now the system says something different – did we do it wrong before?’ They find that difficult. It takes a while for employees to realize that AI can help them make better decisions. Once they see it, it will be quickly adopted.

At the same time, you shouldn’t blindly rely on AI. A system always has its limits, that’s completely normal. For example, it doesn’t know that today we decide to launch a pink T-shirt campaign in two weeks. Especially in a dynamic and volatile business like fashion, you always have to think for yourself and stay in control.

What’s next for you with AI? What are the biggest opportunities in this area for MS Mode?

MS Mode is currently working on implementing the other Thunderstock modules for collection creation (purchasing) and discounts.

We are starting to optimize the purchasing of our never-out-of-stock (NOS) items. However, the goal is to ultimately implement this for the entire collection. If you put together a collection from the start that is more in line with actual demand, there is less need to correct or reduce later.

And if discounts are necessary, we can use them in a much more targeted manner thanks to AI demand forecasting: at the right time and with the right discount. This leads to higher margins and faster inventory turnover.

Ultimately, we want to close the loop and improve the entire primary process – purchasing, inventory and discounting – using AI. As a retailer, you are always working on these core issues. Now we have technology that can help us with this.

And beyond MS Mode: Where do you see the biggest opportunities for AI in the fashion industry?

Many fashion companies find themselves in the same cycle again and again: If sales drop for a short time – and yes, something always happens in fashion – there is too much product left. Then the sales follow, and before you know it, Black Friday is just around the corner. This means that everyone has to massively reduce prices and retailers have to follow suit, even if business is going well. As a company, we are dragging the industry down and the pressure on margins and profits remains high.

If we use AI more intelligently, we can break this cycle. Then the industry will become more efficient, healthier and ultimately more attractive for customers. So the benefit of AI goes beyond just business results: it can make the entire fashion industry more enjoyable and lucrative.

A tip or advice for others in the field of AI?

AI is still in its early stages in the fashion industry. The technology is developing rapidly, but its application is still in its infancy. It reminds me a little of when the Internet came along: there was a lot of talk about it, but it took years for it to really be applied and commercialized.

If I can give one piece of advice to others, it would be this: just start. Don’t wait until AI is fully developed or you know exactly which programs work well and which don’t. This moment will never come. You just have to get in, get started and experience it for yourself.

What we do today will undoubtedly be considered hopelessly outdated in a few years, or perhaps even much sooner. But that doesn’t matter. The most important thing is that you do it. You only learn by doing.

MS Mode is on a stable course, Mok concludes when asked. “We’re growing in stores and our e-commerce is also continuing to grow. We’re deliberately approaching this in a calm and controlled manner. Our like-for-like sales look good and are growing steadily. In that respect, it’s not particularly exciting, and that actually suits us quite well.”

.

ttn-12