Careers in AI — AI in Retail

This article is part of the ‘Careers in AI’ series. In this article we explore how Artificial Intelligence is used in retail and reveal what to expect if you decide to pursue a career in this intersection.

Authors: Andre Frade

Artificial intelligence is reshaping the retail sector, providing new tools that enable retailers to compete and thrive in overcrowded market spaces.

What is the retail industry about?

The retail industry refers to the economy segment that sells goods and services to the public. The sector comprises a broad range of traditional and online businesses that often buy goods in bulk from wholesalers to resell them to individual consumers in smaller quantities. Retailers may include supermarkets, franchises, department stores and independent shops.

Why consider machine learning in retail?

The recent digital revolution brought up a new generation of shoppers that expect premium shopping experiences at low prices. Nowadays, items can be easily found, compared and bought at a click of a button. The exposure to multiple options is huge and the competition is aggressive. Retailers are forced to act fast and artificial Intelligence has proven to offer the competitive advantage that they need to thrive in overcrowded market spaces.

Artificial intelligence tools gather, integrate and analyse consumer data — demographics, social media activity, purchasing histories, etc — to learn how retailers can best understand and connect to customers. Data driven insights and decisions often lead to improved customer experiences, but also to smarter business operations, fast innovation, and identification of opportunities to grow revenue — all of which help differentiate retailers from their competitors.

The intersection

In the retail space, machine learning technologies are employed with two main focuses. Some tools are dedicated to improving customer experience, whilst others are used to improve business operations.

In a competitive world, retailers work hard to connect with their customers and fulfil their demands and expectations. AI technologies are capable of discerning customer profiles, patterns and preferences. Retailers use this information to understand their crowd and make customer experiences more convenient, personalised and enjoyable. Some of the wide range of strategies and algorithms that are brought together to achieve this goal include personalised advertisement, targeted shopping recommendations and promotions, easy checkouts, and hassle-free returns. Satisfied customers tend to spend more and faster than initially planned.

On the other hand, AI technologies help retailers to run their businesses more efficiently by identifying opportunities to decrease waste and improve revenue. Data driven insights help retailers to improve demand forecasting and inventory management, optimise product placement, make pricing decisions, and design better marketing strategies, such that customers connect with the right products at the right time.

Future

The impact of AI tools in the retail industry is undeniable. Retailers that invest into new solutions are able to differentiate themselves and thrive. However, the dynamic and fast paced nature of the sector brings new pressures to retailers: the competition becomes stronger and customers expect more.

Thus, major retailers are already investing in the new generation of AI powered shopping tools and most seem to focus on personalisation. For example, North Face is helping customers find the perfect coat, Olay uses AI to personalise skincare, and American Eagle is developing the fitting rooms of the future with interactive mirrors that allow shoppers to search for clothing, request sizes and keep track of their basket.

Types of Applications

Demand and competition have been driving the development of new and creative AI powered retail solutions. Below you may find some of the most popular applications, which include demand forecasting, operational optimisation, customer profiling, personalisation, and shopper assistants.

Demand Forecasting
Machine learning algorithms help retailers to combine consumer and competitor data to accurately forecast market shifts and trends in real time. This information helps retailers to be proactive, act fast and prepare. Knowing what to expect ahead of time enables retailers to make optimised business decisions and keep up with the market demands. John Lewis and Lush are examples of businesses acting in this space.

Operational Optimisation
AI-based logistic management tools act on the demand forecasting information to help retailers optimise resource allocation according to market demands. Operational optimisation can happen at different levels, including inventory management, staff distribution, or efficient delivery route planning. Algorithms are often specific to each stage and can be implemented together or in isolation. Ultimately, they provide operational adjustment solutions to ensure that businesses maximise profits and reduce waste whilst delivering exceptional customer service at each step of the way. Zara and Starship Technologies are examples of businesses acting in this space.

Customer Profiling
Advanced AI systems are designed to understand customers by deducing their shopping patterns and preferences without ever explicitly asking for this information. These systems use different customer engagement strategies to acquire information such as demographic data, social media behaviour, or purchase history, to discern consumer profiles and enable retailers to meet consumer’s demands and expectations. Walgreens and Nars are examples of businesses acting in this space.

Personalisation
AI algorithms use consumers patterns and preferences data to help retailers improve the shopping experience through personalisation in both traditional and online shops. Traditional shops often adapt and optimise their in-store product displays, pricing, service and staff distribution, whilst online shops use these insights to deliver personalised online content, tailored product recommendations, or targeted ads, promotions and rewards. YooxMirror and ASOS are examples of businesses acting in this space.

Shopper Assistants
Chat bot assistants are designed to provide quick and painless support, improving customer service and engagement. These systems are prepared to answer common questions and direct consumers to helpful answers and fast solutions. In turn, these bots also collect valuable customer data that can be used to inform future business decisions. H&M and Lidl are examples of businesses acting in this space.

Types of companies

The retail sector is composed of a wide range of businesses, operating in-store and/or online. Big companies have started to grow their own data science teams to develop personalised in-house solutions. However, the high investment required to build a new team is not justified or even prohibitive to some companies, which often outsource their AI solutions from various dedicated businesses, often start-ups.

Examples of retailers with their own data science department:
- Ocado, Tesco, Morrisons
- Amazon, Netflix, Spotify
- FarFetch, H&M, Nike

Examples of AI dedicated businesses:
- Bossa Nova Robotics
- Pypestream
- Deep North
- Sentiance

Types of Jobs

Market Research Analyst
A Market Research Analyst integrates and analyses vast amounts of customer and market data — economics, demographics, behavioural, emotional, to derive insight and build optimised business solutions.

Customer Analytics Data Scientist
A Data Scientist in Customer Analytics explores and analyses shopper data to unlock value across a number of business areas. Their responsibilities include developing a solid understanding of the business products and marketing strategies, to then propose solutions that support the business needs.

Demand Forecasting Data Scientist
A Data Scientist in Demand Forecasting is responsible for translating demand forecasting business concerns into analytical questions, to then conduct research using statistical and machine learning methods and derive useful insight and feedback.

Logistics Performance Analyst
A Logistics Performance Analyst identifies areas for improvement in both operational and systems driven processes, as well as provide and implement improvement solutions across the chain.

Example of interview questions

Examples of Non-Technical questions:
What is the worst mistake you ever made?
What is peculiar about you?
How would you solve problems if you were from Mars?
Tell me about a project that went beyond your scope of work.
How would you tell a customer what Wi-Fi is?
Tell me about a time you had to overstep management to get your point of view across.

Examples of Technical questions:
How to estimate the parameters in a uniform distribution for a given data set?
Given n samples from a uniform distribution [0, d], how to estimate d?
Given a sample set of tables, write a sql query to get a summary metric from those tables.
How can we deal with extreme values in data?
How would you detect anomalous behaviour on a user account?
Write a function that gives make the cth column of the rth row of pascal’s triangle.

Links & References

https://global.hitachi-solutions.com/blog/ai-in-retail
https://emerj.com/ai-sector-overviews/artificial-intelligence-retail/
https://www.forbes.com/sites/blakemorgan/2019/03/04/the-20-best-examples-of-using-artificial-intelligence-for-retail-experiences/?sh=44eeeb0e4466
https://www.intel.co.uk/content/www/uk/en/retail/solutions/ai-in-retail.html#:~:text=AI%20is%20enabling%20retail%20systems,can%20provide%20additional%20business%20insights.
https://vue.ai/blog/intelligent-retail-automation/ai-in-retail-for-2020/
https://www.ai-startups.org/top/retail/
https://www.insider-trends.com/34-of-the-best-ai-retail-applications-right-now/
https://medium.com/acing-ai/spotify-data-science-interview-questions-bc7c32c7f637
https://medium.com/@horizons/top-interview-questions-at-amazon-5c6bcebfdb28