A Microsoft Build 2025 session intended to cover AI security turned chaotic when a protest interrupted a live presentation by Neta Haiby, Microsoft’s head of AI security, and Sarah Bird, the company’s head of responsible AI. Amid the disruption, Haiby switched her screen share to Microsoft Teams, inadvertently revealing a confidential internal conversation about Walmart’s upcoming AI deployments.
The leak exposed details of how Microsoft is working closely with Walmart to integrate its Entra Web identity platform and AI Gateway tools into the retailer’s internal systems, plans that had not been made public.
Walmart, one of Microsoft’s largest enterprise clients, is reportedly preparing to expand its use of Microsoft’s AI infrastructure well beyond Azure OpenAI. A message shared by Microsoft principal cloud solution architect Leigh Samons described Walmart as “ready to ROCK AND ROLL with Entra Web and AI Gateway.”
The chat also referenced a proprietary Walmart-built AI assistant that was described as “overly powerful and needs guardrails,” a tool called MyAssistant, which uses Walmart’s private data and runs on Microsoft’s Azure OpenAI Service. Messages from a “distinguished” Walmart AI engineer also praised Microsoft for being “WAY ahead of Google with AI security,” reinforcing Microsoft’s competitive advantage in AI governance.
Why It Matters: Walmart’s deepening investment in Microsoft’s AI and security tools indicates a serious commitment to responsible AI deployment across its workforce. It also spotlights the challenges tech companies face in balancing information security with public-facing events, especially amid growing scrutiny over the ethical use of AI in defense and geopolitics.
- Walmart to Deploy Entra Web and AI Gateway: The internal chat revealed that Walmart is preparing to roll out Microsoft’s Entra Web, an identity and access management platform, along with the AI Gateway, which provides secure access to generative AI tools across enterprise systems. This indicates Walmart is building a tightly governed AI infrastructure, likely intended to scale responsibly within its massive operations. These tools are designed to enforce security policies, manage authentication, and gate AI access for compliance and risk mitigation.
- Confidential Messages Exposed: Shortly after the protest ended, Neta Haiby mistakenly revealed internal Microsoft Teams messages while sharing her screen. The chat included a message from a Microsoft cloud architect stating that “Walmart is ready to rock and roll” with Entra and AI Gateway. This unintentional leak revealed Walmart’s plans to deepen its integration with Microsoft’s enterprise AI tools.
- Walmart’s AI Commitment: One message quoted a Walmart AI engineer praising Microsoft’s lead over Google in AI security, noting their excitement about the collaboration. Walmart already uses Microsoft’s Azure OpenAI services, and this suggests further investment in Entra and AI Gateway. The messages reflect Walmart’s growing reliance on Microsoft’s infrastructure for secure, enterprise-grade AI deployments.
- Praise for Microsoft’s Lead in AI Security: One Walmart AI engineer reportedly said, “Microsoft is WAY ahead of Google with AI security,” emphasizing the retailer’s confidence in Microsoft’s AI governance approach. This points to security as a differentiator in vendor selection, a key factor for a company handling vast consumer and operational data. It also suggests that AI security, rather than raw model performance, is central to Walmart’s AI vendor strategy.
- Walmart as a Showcase Client: The leaked excitement from Microsoft engineers suggests Walmart may be a showcase client for these tools. If Walmart’s deployment proves successful, it could serve as a model for other large enterprises considering Microsoft for secure AI adoption. The public nature of the leak, however, may complicate how Microsoft positions this relationship moving forward.
Walmart AI details leaked during Microsoft Build conference – CNBC
