AI use in the retail sector has seen a significant increase, posing both opportunities and challenges for data security and oversight. As artificial intelligence becomes more integrated into retail operations, businesses are facing difficulties in tracking and managing the associated risks, according to a recent analysis by the security firm Netskope.
Growing AI Integration and Security Oversight
Over the past year, the retail industry has rapidly transitioned from experimenting with AI to broadly adopting the technology. Netskope’s recent report, released on Tuesday, identifies this trend and offers recommendations for enhancing control and reducing exposure to sensitive data.
Many retailers are now centralising the management of AI tools. The percentage of retail employees using unregulated AI tools has dropped significantly from 70% to 44%. In contrast, the use of officially sanctioned AI tools has increased from 40% to 73%, leaving two-thirds of retail workers utilising standalone AI applications.
Challenges with AI Tool Management
Despite improvement in managing AI use, the report notes ongoing risks due to the broader deployment of AI functionalities within software. Nearly all retail employees engage with software embedding AI capabilities, and 90% report handling AI tools trained on customer data. This unregulated use elevates the threat of exposing sensitive information, as highlighted by Netskope researchers.
The report also shows significant compliance risks, with 56% of AI-related data policy breaches involving federally or state-regulated data, 20% relating to source code exposure, and 16% involving password or API key leaks. Such breaches could have far-reaching impacts on retailers, particularly in protecting their market reputation.
Agentic AI and Threats from Remote Access
Rising use of agentic AI presents additional security concerns. The report describes a 400% rise in retail AI agents interacting with remote Model Context Protocol (MCP) servers, complicating the ability of companies to monitor data movements across these connections.
Netskope’s research suggests that these connections could enable the uncontrolled distribution of business or sensitive data, making improved visibility and management crucial for retailers.
Phishing Threats Targeting AI Users
Phishing schemes targeting AI users are also on the rise. After a dip between May 2025 and December 2025, phishing using AI-related traps surged again early in 2026. By March, approximately 100 in every 100,000 users were targeted by these phishing attempts, highlighting an ongoing risk as employees continuously explore new AI solutions.
To guard against potential security failures linked to AI tools, Netskope advises retailers to act proactively. Suggested measures include blocking unnecessary applications, scrutinising web traffic, and employing data-loss-prevention strategies to detect and mitigate the unauthorised transfer of sensitive information to ungoverned AI instruments.
This analysis by Netskope is based on research conducted from 1 July 2025 to 30 July 2026, highlighting the urgency for retailers to address these emerging challenges associated with AI deployment.
Reporting based on coverage first published by retaildive.com.