AI Talk: The retail revolution - tech meets the next-gen shopper  

Views on news 

Google launched the Universal Commerce Protocol (UCP), an open standard designed to enable agentic commerce, where AI-driven shopping agents can complete tasks end to end, from product discovery to checkout and post-purchase management. UCP provides a secure, standardised method for AI agents to connect with business backends across the commerce ecosystem. Businesses expose capabilities, which can be extended with features such as discounts, and agents dynamically discover available services and payment options through business profiles. UCP will be a gamechanger for large commerce players. It’s not a single protocol but a stack built upon a couple of layers – for example, the Agent-to-Agent protocol. In the long term, it has the potential to level the playing field between large and smaller players in the e-commerce arena.   

 

Personalisation versus consumer data protection and the best RoI 

AI adoption is still in the FOMO phase, with many companies throwing AI at problems which could be solved with less advanced technologies too, and some of the younger generations are now even pushing back on AI-driven solutions. Meanwhile, big brands are pushing their products into the AI world. AI is a fast-paced space, the latest technologies now being context graphs and agentic AI, and retail is in the phase of figuring out what the best technological fits are for existing problems. To tackle the ethical aspects of personalisation, analytics must use anonymised data, while the fundamentals of data governance still do apply. To stay on the safe side, AI applications should be deployed internally first, while existing good practices should be extended to AI as well. Shoppers feel more relaxed if they own their own data while using personalisation tools, rather than letting a retailer do it for them. Consumers, however, must consider the trade-offs that come with taking advantage of AI-driven personalisation just like they do in other contexts along their shopping journeys. If providing personal data can improve the shopping experience considerably, they may not act on their privacy concerns during a 30 second purchase.  

 

When LLMs generate responses for market research, these tend to be average responses. To get more individualised ones, the AI customer personas must live in a number of different models, the results of which then can be aggregated and subjected to quantitative analysis. When choosing between different AI-empowered capabilities, such as dynamic pricing or agentic shopping, retailers should start with assessing impact and business value. The areas where AI brings the biggest RoI include idea generation – while interviews with customers brought up 10-20 ideas, AI can come up with a hundred. In the next phase, researchers can leverage AI to whittle down those ideas into 10. AI can also identify patterns that humans can’t or assess what impact a politician’s announcement can have on sales locally or even globally and, thus, enable demand planning and supply chain optimisation.  

 

The panel’s advice 

• 75% of GenZ says they don’t engage with brands that don’t offer AI-driven personalisation. 
• Even if you do AI-driven research, never leave humans out. 
• State-of-the-art AI technology is easier to deploy and integrate with other systems than previous models.  
• AI and data literacy is non-negotiable.  

To listen to the Unriveted podcast with Martin, click here. 

To learn more about data governance from Amit, go to www.linkedin.com/in/amitshivpuja