AI can boost e-commerce, but concerns over cost and privacy remain

SINGAPORE – Artificial intelligence (AI) can boost a firm’s e-commerce capabilities, but bosses and industry experts warn that challenges remain.

Their concerns around cost, biases, security and privacy highlight the need for companies to carefully weigh the pros and cons of adopting AI within their e-commerce applications. 

SGInnovate executive director of investments Tong Hsien-Hui noted that AI can be used within multiple applications in e-commerce, such as using machine learning to optimise inventory control and drive recommendation engines.

Large language models could also run chatbots to improve customer experience.

“Given the competitiveness of the e-commerce market, retailers will likely welcome or at least seriously consider any solutions that can improve their margins, especially in price-sensitive sectors such as fast-moving consumer goods,” Mr Tong noted.

“In such sectors, as one organisation successfully finds ways to optimise their costs through the use of AI tools, others will be forced to play catch-up or seek other solutions that help them achieve similar improvements in order to maintain competitiveness.”

However, Mr Tong pointed out that there are limitations associated with AI tools that firms will need to understand as they seek to implement them in their business. 

For example, companies with e-commerce capabilities may seek to improve customer targeting and product recommendations by using large data models, but they will also need to balance these benefits against the risk of increasing bias in their datasets.

“Organisations that intend to harness AI heavily in their operations will need to take into account the long-term costs of maintaining a team that can continuously upgrade or maintain the tools and platforms required to extract value from their data models,” said Mr Tong. 

Mr Aaron Tan, co-founder and chief executive of Carro, a unicorn start-up dealing with used cars, said AI can help to determine pricing more intelligently and detect any anomalies in the vehicles it sells. The company has been using AI since its incorporation in 2015. 

“AI helps reduce inventory ageing and improve overall operational efficiency across all functions… With the help of AI and technology, the same salesman can be two to three times more productive,” added Mr Tan. 

However, he pointed out that AI presents its own set of challenges, including high computing costs and the expense associated with hiring developers to work on AI models, noting: “Unless the business is at scale, it’s hard to cover the cost for AI.”

A recent survey by customer relationship platform Salesforce found that 80 per cent of companies in Singapore with e-commerce capabilities, such as retail and automotive firms, are already using AI in some shape or form.

It also noted that 42 per cent have already implemented the technology, while 38 per cent are still experimenting. An additional 20 per cent are still evaluating if and how AI can fit into their operations. 

The survey also found that only a small percentage of customers globally fully trust companies to responsibly use their data and employ AI ethically.

Salesforce Asean senior vice-president and general manager Sujith Abraham noted that AI can help companies become more customer-centric by improving product discovery, personalisation and brand loyalty. 

“Our customers in the region are saying that their sales are improving through AI-powered product recommendations and enhancements in the overall customer experience,” said Mr Abraham.

“However, security and trust are top concerns when it comes to AI.

“Businesses will need to tap trusted technology with built-in ethical and responsible guidelines to gain trust from consumers and shareholders.”

Statista data notes that Singapore has one of the world’s fastest-growing e-commerce industries with a projected revenue growth of US$3.5 billion (S$4.7 billion) between 2023 and 2028. The country’s AI market is projected to reach US$10.39 billion by 2030. 

Singapore

proposed a new governance framework for generative AI

on Jan 16 and is now seeking international feedback.

The initiative – termed the Model AI Governance Framework for Generative AI – was developed by the AI Verify Foundation and Infocomm Media Development Authority, and builds on an existing framework from 2019 that covers only traditional AI.

It is expected to be finalised in mid-2024.