Almost six in 10 retailers plan to adopt artificial intelligence (AI), machine learning (ML) and computer vision (CV) technologies within the next year, according to a survey of global retail directors published this year.
The AI in retail study from retail tech hardware company Honeywell, released in August, sought the views of 1,000 industry leaders across the US and EMEA. It found that 38% of respondents are already using these technologies for select use cases or regions, 35% are using them on a larger scale, and 24% are in a pilot phase or in discussions.
Notably, only 3% said they were not using these technologies at all, suggesting the hype around the emergence of AI and similar tech is warranted to an extent. Almost half (48%) of respondents identified AI, ML, and CV as the top technologies expected to have a significant impact on the retail industry over the next three to five years.
Moreover, consumers seem open to AI’s usage by retailers – as long as it helps drive better shopping experiences. A separate survey by e-tail trade association IMRG, in conjunction with global software company Adobe Commerce, shows that people appreciate online retailers that offer personalised customer experiences, and they are generally welcoming of AI if it provides benefits such as relevant search and product recommendations.
According to that research, which questioned 1,000 UK adults and was published in August, 62% of customers said personalised content is an important part of their online experience and 75% said they are more loyal to brands that are strongly personalised to their interests.
Understanding the benefits of AI
It is becoming apparent AI can bring retailers benefits in all areas of their operations. The Honeywell survey shows the main reasons retailers would deploy new technologies such as AI were improving customer experience (59%), driving greater productivity (49%), and achieving cost efficiencies or a return on investment (44%).
Survey respondents predict AI, ML, and CV will bring the greatest value to four key functions in retail. This golden quartet for AI usefulness, in the eyes of industry leaders, is automating day-to-day tasks such as picking and scheduling; supporting customer service, including live chat; creating targeted customer marketing campaigns; and improving inventory management.
Paula Bobbett, chief digital officer at health and beauty retailer Boots, says the AI benefits for retailers fall into three camps: customer, colleague, and process.
“For colleagues, it should make their lives easier – for example, by quickly allowing them to ideate, automating copy and content generation, or even enabling them to more easily find internal answers to their questions,” she notes.
Answers to questions such as “What’s the HR policy on x?” and “How were my sales last week and what drove them?” could be fast-tracked using AI, Bobbett notes.
As for processes, AI should be able to automate time-consuming processes in retail or simplify supply chains, she adds.
Honeywell’s research suggests new tech will complement the future workforce, with most retailers stating AI, ML, and CV will be used primarily as tools to augment and maximise their workforce, rather than to replace employees. Only 7% of those surveyed said their primary purpose for these solutions would be to reduce human labour.
The new technologies, Honeywell indicates, can enable better utilisation of the workforce through predictive analytics, which in the best cases will lead to improved job satisfaction and more time for people to focus on higher-value tasks.
Retailers questioned said the three primary barriers to widespread adoption of AI and other emerging and potentially transformative technologies are budget restrictions (39%), difficulty in demonstrating business value (29%), and lack of internal expertise to maintain the technology (21%).
Marcel Borlin, chief technology officer at department store group Harvey Nichols, says lots of the conversation around AI right now could be categorised as hype. “But it’s a tech that will mature pretty quickly and there will be user cases where AI can be used in all retail,” he adds.
Borlin says AI is and will be implemented by volume retailers where there is scale and the breadth of data, saying “that is where AI can make a big difference”.
“Weather models for supermarkets, for example. AI will get better at predicting the weather to aid demand forecasting,” he says, adding that retailers will be able to better make decisions over stock for specific seasonal occasions.
Global grocery group Carrefour was one of the first retailers to publicly detail its work deploying digital features based around OpenAI’s generative AI ChatGPT technology. One is an advice robot for shopping on its French site, which includes description sheets for Carrefour brand products and support for purchasing procedures.
Carrefour, which brought the features to life in collaboration with OpenAI partners Bain & Company and Microsoft, also launched Hopla, a chatbot based on ChatGPT, which will be integrated into its French site. The natural-language AI can help consumers with their daily shopping based on budget, food constraints or menu ideas.
Internally, the retail group also revealed it is using generative AI for internal purchasing processes. Development of this initially started with the non-retail purchasing division, helping with tasks such as drafting invitations to tender and analysing quotes.
“AI has an application in the supply chain, especially global ones which are very complicated and need to be as efficient as possible from a cost point of view and environmental perspective,” Borlin remarks. “AI will be able to help create balanced supply chains that have the best impact on the planet but also do not cost a fortune.
“The world is becoming more complex, so you need more complex mathematic models and more complex computing power. Therefore, to make the best use of all the data and optimise everything from supply chain to the customer offer through personalisation is all possible and so there should be an appetite for it.”
The customer perspective
Bobbett argues there is no one-size-fits-all approach to using AI in consumer interactions. “For the customer, AI should make it easier to engage with a brand digitally, get more advice and curation of product finding, and move e-commerce from purely transactional to conversational,” she says.
“But I think it depends on the customer – not all customers are happy to share their information and they want their experience to be private. We have to be careful to make sure that, as a retailer, we have put the right guardrails in place to ensure that the recommendations being made are right – this becomes particularly important in the healthcare space and where you are dealing with customers who may be vulnerable.”
Interestingly, everyone Computer Weekly spoke to for this report emphasised caution when implementing AI.
Borlin urges retailers to “keep an eye on the regulation” around AI: “Before you know it, things could go too far and then it becomes creepy. I’m not so worried about the Terminator scenario – it’s more about the stalking when consumers share too much data and interactions become uncomfortable.
“Technologists and retailers might think it’s great, but people might get creeped out. That’s where we need to tread a very fine line when it comes to AI and personalisation.”
Technologists and retailers might think [AI’s] great, but people might get creeped out…we need to tread a very fine line when it comes to AI and personalisation Marcel Borlin, Harvey Nichols
As AI matures, Borlin estimates it will become more useful predicting consumer trends: “But that is said with caution, because there is still some magic that needs to happen in fashion, for example, due to the emotional engagement involved with shopping for it.”
John Bovill, executive consultant, suggests retail must be laser-focused on the end consumer at all times and should be wary of giving too much control to AI around customer service, which is where retailers are well placed to showcase their brand differentiation.
It is important retailers are cautious with any AI investment as they need to remember the “human factor” which is fundamental to how they operate, he explains.
“Retailers need to remember at the end of any AI-based interaction there is a human there. There has to be a high level of trust and you have to be sensitive, reliable and respectful with the customer’s data. I don’t believe you’re going to be a hero in being bleeding edge in this – you have to tread carefully.”
His comments are supported by the IMRG-Adobe study, which found 18% of customers find personalisation and AI “a bit intrusive/creepy”. The research said customers would mostly feel sceptical (34%) if online retail chatbots became as advanced as ChatGPT.
In addition, 51% of customers are uncomfortable with their gadgets listening to their voice or conversations to make product recommendations.
Andy Mulcahy, strategy and insight director at IMRG, notes: “The consistent thread that comes out of research such as this is that, generally speaking, how comfortable people are with AI correlates with the extent to which it enhances the quality of the experience they get.
“It follows that the anxieties that people may have over how sophisticated and potentially intrusive technology is becoming can be managed by really focusing on the benefits it brings to them and by avoiding the trap of ‘doing AI’ for the sake of it, which many businesses seem sure to fall into.”
Bovill recognises AI is already “all around us anyway in everything we do”. From online search to voice-recognition technology – everything of this nature has AI or ML in some form, he explains, but “we have to keep an eye on it”.
“AI will have a significant impact and it [already has it] today,” says Bovill. “But will it fundamentally change retail? I’m not sure. Retailers tend to have great people selling their products on human values, and if you lose touch with why a brand exists, it’s dangerous. You have to be careful. If some elements of AI create a better experience for customers, clearly retailers then need to consider using it.”