Why retailers should pick up the pace of AI adoption

Analysts said they will run the risk of lagging in the market and suffering higher costs.

Retailers these days cannot just turn a blind eye to the power of artificial intelligence (AI). Analysts caution that neglecting this transformation as an imperative not only risks customer attrition to more tech-savvy competitors, but also the burden of increased operational costs.

Lazada, one of the leading e-commerce platforms in Southeast Asia, is no stranger to AI adoption as it relied on the technology even before its boom, employing AI image-based searches as early as 2019.

“With the mission to accelerate progress through commerce and technology, we have had a long history of leveraging AI and technology to improve the online shopping journey,” Howard Wang, chief technology officer at Lazada Group, told Retail Asia.

In May 2023, the company introduced OpenAI ChatGPT-powered LazzieChat, the first chatbot of its kind in the region that is available in English, Bahasa Indonesia, and Tagalog for the Philippines. This technology was designed to answer customers’ shopping queries to provide them with a personalised experience.

“Our approach to AI is reflective of our dedication to continuous improvement and tech advancement, as we seek new ways to deliver a superior online shopping experience for all our users. We continue to work closely with AI experts and developers to apply the potential of AI to unlock a new era of retail innovation,” Wang said.

Analysts assert that with this proven track record of enhancing retail operations and bolstering business performance, the urgency of implementing AI strategies is clear and undeniable for the industry.

Risk of losing out

Anson Bailey, head of consumer and retail at KPMG Asia Pacific, said retailers may “lose out” if they do not keep the pace in AI adoption, noting that the use of the technology in the marketing aspect often leads to a winner-takes-all situation. This is because the first to implement a fun chatbot would drive the most online traffic as customers try the feature.

If they do not leverage the machine learning and analytics side of AI, they also risk losing their market share to competitors as they do not have access to insights that can be used to align product offerings and experience to customers.

“Retailers today can’t afford to ignore AI nor can they sit idly by, whilst their competitors and new economy players gather momentum in the marketplace,” Bailey said when interviewed by Retail Asia.

Also speaking to the magazine, EY Asean Consumer Products and Retail Leader Olivier Gergele pointed out how companies can automate manual tasks and enhance efficiency and accuracy as well as customer experience through AI technology.

By doing so, businesses can focus more on assigning their talents to strategic responsibilities. 

Gergele said those who fail to integrate AI “miss out on opportunities to optimise the business in the areas of inventory management, supply chain logistics, and pricing strategies, leading to inefficiencies and higher operational costs.”

He also highlighted the benefits of valuable consumer insights that can be driven by AI, which helps in predicting demands and trends and providing information for operational decisions.

“AI-driven personalisation and recommendations can boost sales and consumer loyalty, and with heightening consumer expectations for tailored experiences, these missed opportunities may eventually erode the companies’ competitive advantage and potentially, market share,” the EY retail specialist said.

Use cases

Retailers will find AI most crucial in increasing efficiency in their operations. “AI allows retail companies to concentrate their staff attention on the most pressing and complex issues, and automate repetitive, mundane tasks. This helps their bottom line and frees up time and manpower for new operational projects,” said KPMG’s Bailey.

For example, some retailers automate store traffic analysis by feeding camera data to AI and machine learning algorithms. Japan’s Aeon Retail uses AI video analysis to identify shopper gender and age alongside their shopping patterns in-store.

E-commerce giants are also launching generative AI models to aid shoppers in product search and improve customer experience. Aside from integrating ChatGPT, some developed their own AI language models, such as Alibaba’s Tongyi Qianwen.

Aside from the operational aspect, AI can also be utilised on the front end, EY’s Gergele asserted.

He said AI algorithms allow real-time price adjustments based on demand and conditions. It can boost the effectiveness of marketing and advertising campaigns by analysing consumer behaviour to generate targeted campaigns. The chatbots will also provide 24/7 customer support.

“Given that it is easier and cheaper to retain existing consumers than to acquire new ones, companies can tap into AI to design digital marketing strategies that drive deep, long-term and enduring relationships. They then not just retain consumers, but also generate additional sales from them through cross-selling and up-selling,” Gergele said. 

At Lazada, Wang said they use the technology to provide translation to reach more customers, enabling seamless interactions between buyers and sellers.

Wang also said that the AI-powered search recommendations comprise half of Lazada’s total user transaction and platform as it provides real-time personalisation. It is also beneficial to its logistics network because it provides delivery riders with more efficient routes.

In September this year, Lazada also integrated AI and augmented reality applications for skin test technology where consumers could make skin diagnosis and analysis through their phones and get recommendations for relevant products.

Lazada’s Creative Centre for brands and sellers also leverages AI-generated content technology to improve the quality of their creatives, therefore saving them time and effort, Wang said.

Challenges in AI adoption

Whilst AI is rapidly emerging, Gergele noted that the technology is still at a “nascent stage.” “Companies need to understand the latest developments and best practices, and evaluate the most suitable AI solutions that support their business needs and goals,” he said.

“AI systems rely on large amounts of data — and therefore adoption brings with it the associated risks of cyber and data breaches. Further, the data used to train the AI systems may have bias, which can impact the AI-generated output,” he added. 

Despite these advantages, there remain challenges in fully integrating AI into retail operations.

Bailey said concerns faced by businesses were around customer privacy and compliance with legislation such as the Chinese equivalent of the General Data Protection Regulation or Personal Information Protection Law.

“Large businesses are willing to invest in technologies that can highlight new customer insights, but their priority is with customer privacy,” he said.

Business leaders also wait for their staff to get set on implementing and using AI to officially launch it. However, they usually face hurdles in finding sufficient training methods for their employees.

There are also issues of potential displacements due to automation whilst the workforce is upskilling to adapt to AI. Gergele said this calls for a review of talent strategy and investment in training to make them capable of using AI.

Advanced AI systems can also be expensive, especially for small or medium-sized businesses, he said. 

“Notwithstanding that cost is expected to decline with widespread adoption, companies may consider starting small with a few select and targeted solutions to demonstrate successful use cases, before expanding AI adoption organisation-wide,” Gergele told Retail Asia.

Making it work

To integrate AI in retail operations, businesses should first have a clear strategy of what they want with AI-improving inventory management personalisation, amongst others, said Julian Cua, partner at Boston Consulting Group.

With finite resources, retailers should focus on a few strong use cases, Cua said. Otherwise, they run the risk of having many small things but have no full implementation of some big ones.

There are also several risks in the implementation of the technology itself, one of which is ensuring having the tech stack that will consume and process all the data in a cost-efficient way. 

“The more data you consume, the more it costs. So you have to balance that out and be smart on how efficiently you’re actually engineering your data and that goes to how you're basically building your data platform or data hub,” Cua told Retail Asia.

Getting data is also vital as AI would not exist without it, Cua said. Retailers should be clear about the data they want to get and how to capture them, noting that it is difficult in some markets such as the Philippines as a lot of transactions there happen offline and are recorded on paper.

Aside from ensuring the data privacy of consumers, Cua said retailers should also focus on the people and processes around AI as they comprise 70% of the success factor of AI implementation.

“We’ve seen many companies invest in the best, expensive solutions out there in the world. But if it’s not implemented well, with the right people, with the right change management, with the right changes in the processes that enable those outcomes, then it’s not going to realise the value,” Cua said

 “The biggest key variable in digital transformation in data transformation and implementing AI use cases are the people and process and organisation that is around it,” he said.