Dive Brief:

• In the retail sector, artificial intelligence (AI) represents a $300 billion opportunity for those companies able to scale deployments across the entire value chain. However, only 1% of retailer uses cases have reached this level so far, according to a recent survey of 400 global retailers implementing AI from the Capgemini Research Institute.
• Of the retailers surveyed, 28% are deploying AI in some form today, up from 17% in 2017 and 4% in 2016. Those retailers said AI is creating jobs rather than replacing workers, and any job cuts have been 25 or fewer positions.
• While most uses cases focus on customer relations and sales, there is a 7.6% or $144 billion savings opportunity from supply chain through improved efficiency in routing, warehousing, returns management and procurement.

Dive Insight:

Retailers are at different stages of maturing in implementing AI in various use cases, as reported in the study.

The group of 400 retailers represented 23% of the global retail market by revenue, and additional public data from the largest 250 retailers by revenue was also included. Large retailers with $10 billion or more in revenue were the most active in AI implementation.

"Our research shows a clear imbalance of organizations prioritizing cost, data and ROI when deploying AI, with only a small minority considering the customer pain points also," Kees Jacobs, vice president, global consumer products and retail sector at Capgemini, said in a press release. "These two factors need to be given equal weighting if long-term AI growth with all of the benefits it brings, is to be achieved."

Retailers were eight times more likely to be working on complex, higher return projects rather than simple, customer-centric deployments. Top drivers behind AI implementations include cost (62%) and ROI (59%), while customer experience (10%) and known customer pain points (7%) rank lower on the list of priorities. Fewer than one in five use cases are deployed in areas such as distribution and logistics or in-store operations.

Although operations represent only 26% of AI use cases today, they are among the most profitable. Leading examples included using AI for procurement (averaging 7.9% ROI), applying image-detection-led algorithms that detect in-store pilferage (7.9%) and optimizing supply chain route plans (7.6%).

Major retailers such as U.K.-based Tesco, China’s JD.com and Alibaba, have experimented with AI influenced routing plans in the supply chain. Flipkart, an Indian pure-play online retailer now owned by Walmart, uses machine learning to overcome the challenges of India’s unstructured postal address systems, resolving inconsistencies with a 98% accuracy rate.

Robots in Ocado’s warehouses not only pick and pack autonomously but also collaborate to fulfill a typical 50-item order in minutes. Kroger will use Ocado's technology in new AI-powered warehouses.

German retailer OTTO uses AI to analyze 300 billion past transactions across 200 variables to predict purchasing patterns for the next 30 days. The system has reduced returns by more than two million items per year due to a 90% accuracy rate in its predictions.

Using AI to transform supply chains ultimately improves the customer experience, the report said. Retailers who want to compete will use AI in both operational and customer-facing functions.