EY Deploys Agentic AI Across 130,000 Auditors in 150+ Countries

EY announces a global deployment of enterprise-scale agentic AI across its entire 130,000-person Assurance workforce conducting 160,000 audits in over 150 countries, representing one of the largest production rollouts of multi-agent AI systems. The framework, built on Microsoft Azure and embedded in EY Canvas which processes 1.4 trillion journal entry lines annually, is backed by multibillion-dollar investments and aims for full AI-supported audits by 2028. Meanwhile, OX Security disclosed a critical vulnerability in Anthropic's Model Context Protocol affecting 150+ million downloads and up to 200,000 servers, with Anthropic declining to fix the architecture and instead pushing responsibility to developers. Separately, Canva unveiled AI 2.0 with conversational design capabilities and proprietary models 7 times faster and 30 times cheaper than frontier alternatives, rolling out to 1 million of its 265 million monthly users.

EY's Global Deployment of Agentic AI Across 130,000 Auditors

• ▪EY is part of the inaugural class of Microsoft's Frontier Firm AI Initiative alongside 13 other organizations
• ▪EY expects full end-to-end AI-supported audits by 2028
• ▪EY is running a global training program throughout 2026 for its agentic AI deployment
• ▪EY announced a global rollout of enterprise-scale agentic AI across 130,000 professionals in its Assurance workforce
• ▪EY Canvas processes over 1.4 trillion lines of journal entry data per year
• ▪EY's Assurance workforce conducts 160,000 audits in more than 150 countries

Critical Security Vulnerability in Model Context Protocol (MCP)

• ▪At least 10 high- and critical-severity CVEs exist across tools like Cursor, VS Code, Claude Code, and Gemini-CLI related to the Model Context Protocol vulnerability
• ▪Up to 200,000 servers are vulnerable to the Model Context Protocol security flaw
• ▪The Model Context Protocol has 150 million or more downloads
• ▪OX Security's team successfully compromised 9 of 11 MCP marketplaces during testing

Canva's AI 2.0 Platform Transformation and Agentic Workflow Integration

• ▪Canva AI 2.0 supports conversational design from natural language prompts
• ▪Canva AI 2.0 represents the company's biggest product overhaul since launching in 2013
• ▪Canva AI 2.0 is rolling out to its first 1 million users
• ▪Canva AI 2.0 includes connectors to Slack, Notion, Zoom, Gmail, and Google Calendar
• ▪Canva's proprietary AI models are 7 times faster than comparable frontier alternatives
• ▪Canva AI 2.0 includes a Memory Library that retains brand preferences across sessions

Perspective of OX Security researchers

• ▪OX Security researchers demonstrated the severity of the Model Context Protocol flaw by successfully compromising 9 of 11 MCP marketplaces during testing

Perspective of Canva

• ▪Canva emphasizes its proprietary AI models' 7 times faster speed as a competitive advantage over frontier alternatives
• ▪Canva highlights its proprietary AI models' 30 times lower cost as a key differentiator from frontier alternatives
• ▪Canva frames AI 2.0 as its biggest product overhaul since the company launched in 2013