How AI is transforming the retail landscape

The promise of Generative Artificial Intelligence (GenAI) in retail is as large – possibly larger – than the emergence of the Internet back in the 1990s.

However, it will also come with its own challenges. In this session, Deborah Weinswig, CEO and Founder of Coresight Research and Jon Reily, President, McFadyen Digital, discussed what can work and what might not work, for global retailers with this new technology.

From personalisation at scale to privacy, they discussed what trends retailers need to be watching for when it comes to how best to use GenAI in their businesses, and perhaps more importantly when not to.

Jon spoke about the practical impacts of AI in a retail setting.

“While many think about the creative elements of AI, such as humanoid-style chatbots, AI can simplify and automate mundane daily tasks. AI autonomous inventory robots can catch pricing errors on shelf talker or help facilitate surge pricing.”

Unfortunately, in their early stages, chatbots were a bit frustrating for many customers, equipped with too few phrases and too little helpful information.

“Retailers are investing in smarter conversational AI. AI that pulls history into the conversation, while also using emotional recognition tools to better understand the context of a customer’s inquiry. If AI senses frustration, it can pass the customer to a real team member,”

“Sephora recently implemented AI to help customers find their perfect make-up. Colour IQ scans a customer’s face and provides personalised recommendations for foundation or concealer, while Lip IQ does the same to help find the perfect shade of lipstick.”

Deborah spoke about the applications of AI in online shopping.

“You can work out when customers are exiting their online journeys with your brand, then make meaningful changes to things quickly, to mitigate further ‘leakage’.”

Jon queried what retailers were getting wrong and what they needed to be mindful about.

Deborah provided an example of her company once using ‘Copilot’.

An artificial intelligence-powered Copilot provides real-time assistance to provide immediate help to employees in various tasks offering immediate, context-aware support, enhancing employee productivity, and optimizing business operations as an outcome.

“We used Copilot AI to write minutes from meetings, pull together reports and develop agendas. However, as Copilot pulls data from multiple sources, including our company’s chat software, it can make mistakes.”

“We had a team member ask Copilot AI about their bonus, and it accidentally sent them every team member’s salaries and bonus structures.”.

“Copilot created an agenda for a meeting I was planning with my team. It identified five points that were a work in progress from an earlier meeting to be discussed and it pulled reports and chats from multiple sources, but also picked up a private email between two colleagues, and added it to the agenda, essentially saying “Tell bob his graph looks like crap.”

Jon reminded everyone that Gen AI is just a facilitator. You got to check what it is putting out and where it is pulling information from. There is still a need for a ‘human touch’ as it grows.

Jon suggested AI can augment human workers and teams.

“In the past, AI was seen as a marketing personalisation tool, but that is shifting to an efficiency-driven, cost reduction tool.”

“Now I and my team can access decades of our reports almost immediately. AI can pull out what I’m looking for and produce it in a digestible format.”

“Our creative costs can fall. AI can develop adverts with just a few lines and images.” In the past, firms had teams of people with reporting-related duties. Although much of the data collection process has been automated, teams were still losing countless hours on manual data reporting.”

The discussion shifted to ‘synthetic human focus groups’.

Imagine a world where Charmine, a mum living in Brisbane, discusses her daughter’s love for the Socceroos. She shares her experiences of taking her to a game at Suncorp Stadium and how soccer has become an integral part of their lives. But here’s the twist — Charmine isn’t real. She’s a creation of advanced AI software, a synthetic human with a detailed backstory.

“In this rapidly evolving digital landscape, AI characters like Charmaine and others are providing benefits for retailers. Where once we created static ‘personas’, AI personas are now able to engage in discussions about products and brands.”

Importantly, AI isn’t necessarily new, it has always been here, and it’s just getting better and faster.

“Retailers have been using AI and Machine Learning for years to detect underlying trends and patterns. For example, analysing shopping data of individual purchases in a shopping basket to identify what a customer was making for dinner that evening. Then, offering them an associated product, rather than a discount.

Deborah advised that 60% of business today don’t have an AI policy.

“With the surge of AI integration in many sectors, ethical implications associated with its use have become increasingly significant. Dedicated AI ethicists are now essential for businesses to safely navigate this complex landscape, anticipate potential risks and act responsibly.

AI ethicists are becoming responsible for guiding AI development with ethical considerations, ensuring AI systems align with moral and social values and ensuring compliance with emerging regulations and legislation. The aim here is to mitigate risks and potential harm.

It was suggested that the growth and expansion of AI may align with ‘Moore’s Law’. Initially, Moore’s Law predicted that the number of ‘components on a single computer chip’ doubles every two years at minimal cost. This observation has held steady since the early 1960’s.

Jon finally offered a warning about AI, similar to Deborah’s example of AI errors, referred to as ‘AI hallucinations’.

An AI hallucination is a phenomenon where a generative AI chatbot or computer vision tool perceives patterns or objects that are non-existent, and then creates outputs that are nonsensical or altogether inaccurate.

Jon recommends caution. “You only get out what you put in. So what you put into your AI system, needs to be accurate and good quality data.”

SHARE THIS ARTICLE

FURTHER READING

Retail Voice CEO Message: 10 July 2024

Earlier today, the ARA voiced its disappointment on the decision of the NSW Government to ban trading for larger retailers on ANZAC Day – noting that the current approach strikes

Retail Voice CEO Message: 3 July 2024

Creating a more inclusive retail sector is a vital focus for the ARA and last week we were proud to co-host our third annual Diversity, Equity and Inclusion (DE&I) Symposium,