Businesses are becoming tech and data-oriented across industries. They will be able to quickly grasp this concept and live it, which will allow them to meet customer expectations and create more value for their business. It is becoming increasingly important to reimagine the business and to use digital technologies to create new business cultures, experiences, and opportunities.Digital transformation is often portrayed as all about technology. It is not. Digital transformation is a complex process that requires and depends on diversity to be successful. Artificial intelligence (AI), which is enabled by human intelligence's vast talents but also vulnerable to its limitations, is the result.It is essential that organizations and teams make diversity a priority. They should also consider it in broader terms. Diversity is based on three key pillars to me.PeopleArtificial intelligence is a product of people. In fact, artificial intelligence is created by humans. The diversity of the people who make up the AI algorithm development team must reflect the diversity in the population.This includes ensuring women have equal opportunities in technology and AI roles. It also includes all dimensions of gender, race and ethnicity as well as skill set, geography, education, perspective, interests, and other factors. Why? Why?Digital transformation is often portrayed as all about technology. It is not.We have the opportunity to use AI and machine learning together to help propel the future. This starts with diverse teams that reflect the rich diversity and perspectives of our planet.Our digital transformation has been aided by the diversity of our skills, perspectives, experiences, and geographic locations. Levi Strauss & Co.'s growing strategy and AI team does not include only data and machine-learning scientists and engineers. Recently, we tapped employees across the company and set out to train those with little or no experience in statistics and coding. We took employees in retail operations, warehouses and design planning. Then we put them through our first ever machine learning bootcamp. This allowed them to build on their retail skills while also allowing them to master coding and statistics.We didn't limit your options. We simply wanted people who are curious, problem solvers, analytical and willing to try new ways of approaching business problems. Employees who completed the program have new perspectives and valuable business knowledge thanks to their existing retail expertise and machine learning skills. We were able to develop a diverse and talented team through this first-of-its kind initiative in the retail sector.DataMachine learning and AI are only as good or as effective as the data they receive. Data is not limited to structured data, figures and tables. However, data can be digitized.Digital images of jeans and jackets that our company has produced for over 168 years are data. Data also includes customer service conversations, which are only recorded with permissions. Heatmaps of how customers move around our stores are data. Data comes from the reviews of our customers. Everything that can be digitalized today is data. We must think differently about data, and ensure that all data is continuously fed into AI work.Predictive models are based on data from the past in order to predict the future. Because the apparel industry is still at the infancy stages of digital, AI and data adoption, it is difficult to have past data as a reference. Fashion is able to forecast trends and anticipate demand for new products. How can we do this?Data is more important than ever, including images of new products as well as a database of products from previous seasons. Computer vision algorithms are used to identify similarities between the new and old fashion products. This helps us predict future demand for these new products. These applications are able to provide more accurate estimates than intuition or experience, and can be used to supplement previous practices with AI- and data-powered predictions.Levi Strauss & Co. also uses digital images and 3D assets for a virtual experience of how clothes feel. We train neural networks to recognize the subtleties of jean styles such as tapered legs, whisker patterns, and distressed looks. Additionally, we detect the physical properties that affect drapes, folds, and creases. This data was then combined with market data to allow us to tailor our product lines to meet changing consumer demands and desires. We also focus on inclusivity across all demographics. We also use AI to create new styles, while still retaining the creativity of our top-notch designers.Tools and techniquesWe must ensure that there is diversity in the tools and methods we use to create and produce algorithms. Some AI products and systems use classification techniques that can perpetuate gender- or racial biases.One example is that classification methods assume gender is binary. They assign people male or female based upon their physical appearance and stereotypical assumptions. This means that all other forms and gender identities are eliminated. This is a problem and it's up to all of us who work in this area, regardless of industry or company, to avoid bias and develop techniques to capture all the nuances, ranges, and differences in people's lives. We can, for example, remove race from the data in order to make an algorithm race-blind and protect against bias.Open-source tools are used to achieve diversity in our AI products. Open-source libraries and tools are by nature more diverse. They are accessible to everyone, and people from all backgrounds work together to improve and advance them. This enriches their experience and limits bias.Our U.S. Red Tab loyalty programme is an example of this. We don't ask fans to choose a gender nor do we allow the AI system make assumptions when they create their profiles. We ask fans to choose their preferred style (Women or Men, Both, or Dont Know), to help our AI system create tailored shopping experiences for them and provide more relevant product recommendations.Levi Strauss & Co. is transforming its business and the entire industry by embracing diversity of people, data and techniques. Our company's social values have been a legacy for 168 years. We continue to build on that legacy. This is the latest opportunity to build on this legacy and shape fashion's future with diversity in AI.
