About twenty years ago, the retail industry was vastly different. It was defined by manual inventory counts and face-to-face conversations to understand customer preferences.

Nearly half of industry respondents see AI and GenAI as key to enhancing end-to-end supply chain visibility, suggests Deloitte’s Global Retail Outlook report. The same proportion of retailers surveyed believe these technologies will transform personalized product recommendations. While this focus highlights a growing trend, retailers are leveraging AI not just for customer-facing operations, but also to streamline backend processes.

As businesses embrace GenAI, they are also radically shifting their data strategy, pushing AI to the edge. Personalization is ramping up, bringing increased privacy concerns along with it. With data breaches becoming an almost annual tradition, the industry is caught in a challenging dilemma. “Worldwide end-user spending on security and risk management is projected to total $215 billion in 2024, an increase of 14.3% from 2023, according to Gartner®.”

Why Edge?

The shift to edge AI is more than a technological upgrade; nearly three-quarters of consumers now demand tailored interactions from companies, suggests a McKinsey survey. It is about giving customers a unique shopping experience, no matter where they buy – online or at the store, and at the same time, ensuring their information is protected. This change also stems from data privacy concerns, with almost 50% of consumers surveyed globally citing privacy and security as their top concerns when engaging with brands. By processing data locally rather than relying solely on distant cloud servers, retailers can offer personalized interactions across all touchpoints while reducing the risk of breaches and allaying consumer fears about data misuse.

Edge points—from mobile devices and in-store computers to smart shelves and cameras—are becoming increasingly powerful, driven by enhanced computing and memory capabilities. This progression is moving AI processing away from centralised cloud systems and towards these distributed edge points through reduced latency, heightened data security, and hyper-personalization at the customer interface.

Enhancing Store Operations, Security, and Personalized Shopping

The past year has seen a surge in identified use cases and newly developed tools leveraging this cutting-edge technology. In physical stores, edge AI is revolutionizing operations across multiple fronts. Occupancy management systems now provide real-time insights into store traffic patterns, enabling more effective layout designs and planogramming. Computer vision technologies are enhancing promotion efficiency tracking and automating aisle replenishment, ensuring shelves remain stocked and sales opportunities are maximised. Perhaps most crucially, edge AI is tackling the perennial issue of shoplifting, which affects 85% of small business retailers at least once a year, according to Forbes Advisory Survey. The majority (79%) report monthly losses between $500 and $2,500, with 10% suffering even higher losses exceeding $2,500.

Automated self-checkouts and ‘just walk out’ technology are streamlining the purchase process, while AI-driven energy management systems are slashing operational costs. Digital signage is also becoming increasingly personalized, adapting to the demographic profile of passing shoppers. Smart kiosks, both for vending and ordering, are offering tailored recommendations based on individual customer profiles and real-time store promotions.

In customer service, AI is streamlining product returns and warranty claims through swift defect categorization and decision-making. On mobile and web platforms, edge AI is enabling highly personalized targeting without compromising data privacy. Retailer apps can offer timely, personalised recommendations by processing and storing data locally on users’ devices, without transmitting sensitive information to the cloud.

Building Trust Through AI, Data Anonymization

Data anonymization is emerging as a critical safeguard, reconciling the power of personalization with the imperative of privacy protection. In this multi-layered approach to data protection, customer information is predominantly retained on personal devices, significantly reducing the risk associated with cloud storage. Even when stored locally, data undergoes rigorous anonymization processes. This is particularly crucial for sensitive information such as computer vision data, where images and traffic patterns are scrubbed of identifying features before processing.

In many jurisdictions, regulations mandate that retailers can only store data for a specified period—often as short as one to two weeks. Beyond this timeframe, retailers must purge the data or obtain explicit customer consent for continued storage. This ‘destroy-or-consent’ model puts control firmly in the hands of consumers, allowing them to dictate the lifespan of their personal information.

These robust security measures are not just about compliance; they’re about building trust. By implementing stringent anonymization protocols, frequent data refreshes, and consent-based retention policies, retailers are demonstrating a commitment to responsible AI use. This approach allows them to harness the power of personalization and predictive analytics while respecting individual privacy rights.

The Bottomline

The retail landscape is being reshaped by two competing consumer demands: the desire for tailored interactions and the need for data privacy. Edge computing has emerged as a promising tool in this balancing act, offering a way to process data locally and mitigate privacy risks. But it’s not a standalone solution. The industry must be open to adopting multi-layered strategies, combining edge processing with advanced data anonymization to thread the needle between customization and confidentiality, pushing the boundaries of personalized experiences. While it’s a subtle balance to strike, it’s one that’s growing increasingly vital in the AI-powered retail world. For an industry accustomed to constant change, this may be its most significant adaptation yet.

The author is Padmanabhan Venkatesan, Senior Vice President and General Manager – Consumer
Tech, Persistent Systems

Disclaimer: The views expressed are solely of the author and ETCISO does not necessarily subscribe to it. ETCISO shall not be responsible for any damage caused to any person/organization directly or indirectly.