Flying high with AI: dealing with data in a responsible way

In this data-driven age, it would be impossible for human beings alone to manage a family of global retail brands like H&M Group. Here, Errol Koolmeister, Head of AI tech, Architecture & Data Science at H&M Group outlines how we have maximised the superhuman powers of AI to make data part of our corporate culture.

6 Nov, 2019

Who can bear the frustration of not being able to lay their hands on just the fashion item they want when they need it? To Errol Koolmeister, Head of AI tech, Architecture & Data Science at H&M Group, working with AI is one of the most effective ways to enable H&M Group’s growth because growth isn’t just about profits and revenue: it’s about understanding what the customer wants and give it to them, thus becoming more relevant. AI allows us to provide customers with a relevant service, customising the merchandising mix of individual stores and understanding and predicting customer needs.

Speaking at PyCon Sweden 2019 on October 31 at Münchenbrygggeriet in Stockholm, Errol addressed the matter of scaling AI and how data is part of our corporate culture. “I love AI…” he said, explaining how AI brings remarkable efficiency to the matter of doing business.
“It represents valuable insights that allow us to act on the signals we are given and create value by becoming increasingly customer-centric and relevant. At H&M Group we have great people and with AI we can amplify their skills. Have AI recommend what we should focus on and then have people making the decisions. We talk about AI as Amplified Intelligence - mixing the best of man with machine learning.”

"Growth isn’t just about profits and revenue: it’s about understanding what the customer wants."
Errol Koolmeister is Head of AI tech, Architecture & Data Science within the H&M Group Advanced Analytics and AI unit.

A journey determined by data

H&M Group has been a data-driven operation from the start, building a network of physical retail outlets in response to customer demand. “It’s always been about building stores that are relevant,” said Koolmeister. “That’s why, for example, we have five or six in central Manhattan.”

Today, however, online retail accounts for some 14 per cent of the company’s revenue and is increasing. “So we’ve had to adopt a new approach, maximising the use of big data and advanced analytics, and constantly inventing new algorithms to help fulfil our customers’ needs.”

Recognising the value of AI

The AI team at H&M Group have always had strong support from top management. In fact, he says, it was CEO Karl-Johan Persson who in 2015 asked what the company was doing with AI. “By 2016, we’d realised that we needed to do something big!”

That’s when they established a brand-new cloud-based IT infrastructure and set to work on a series of AI and Machine Learning use cases, realising their value and proving that they worked.

H&M Group’s Advanced Analytics and Artificial Intelligence unit now produces fully-fledged analytical software platforms for implementation across the globe.

“The company produces large quantities of garments in a year,” Koolmeister said. “We work on AI use cases that tell us how we can quantify fashion, allocate it and make it personal for each individual customer.”

How we remain future-proof

It is by building “a modular system with a solid backbone” that the AI team can secure a future-proof IT architecture. “This enables us to be agile which is the secret of success – swapping out pieces of the puzzle as necessary and remaining entrepreneurial by adopting new technologies, integrating them and proving their value. Ultimately, it’s about starting small, thinking big and scaling up fast.” In other words, remaining true to H&M Group tradition.

But perhaps most importantly of all, AI gives the customer purchasing options. As Errol Koolmeister puts it: “AI democratises people’s sense of being unique.”

H&M Group uses cookies to give you the best experience on our website. If you continue to use our services, we'll assume that you're happy with this. Find out more about cookies.