Data Governance · 3 min read
Manufacturing AI use jumps from 40% to 67%, Netskope finds
Netskope's November 2026 report: manufacturing AI users grew from 40% to 67%, regulated data drives 57% of violations, and malicious AI links hit 200 per 100,000 users in March.
Key takeaways
- Manufacturing AI users rose from 40% to 67% year over year; managed AI use doubled from 35% to 70%.
- Microsoft Copilot is used by 90% of organizations; Particular Audience is the most-blocked app at 45%.
- Regulated data accounts for 57% of AI-related policy violations in manufacturing.
- Remote MCP connections by non-human entities grew 490%; MCP events grew 600%.
- Malicious AI link interactions reached 200 per 100,000 users in March 2026.

AI adoption among manufacturing users rose from 40% to 67% in a single year, while organization-managed AI application use doubled from 35% to 70%, according to the November 2026 report from Netskope Threat Labs. The study draws on anonymized telemetry from Netskope One customers collected between June 1, 2025 and Aug. 31, 2026, and it documents a sector whose data governance controls have not kept pace with its tooling.
Microsoft Copilot leads adoption at 90% of organizations — a pattern Netskope says distinguishes manufacturing from other industries. Claude Platform follows at 89%, and ChatGPT reaches 76%. Direct use is only part of the footprint: 96% of employees work with applications carrying embedded AI features, and 90% interact with systems that use customer or user data for model training.
What is leaking, and where?
Regulated data accounts for 57% of AI-related data policy violations in manufacturing. Passwords and keys follow at 19%, intellectual property at 16%, and source code at 9%. Most AI use happens inside applications employees already rely on, not in standalone tools, which makes it harder for security teams to track where sensitive information travels.
Remote Model Context Protocol (MCP, a protocol that lets AI agents connect to external tools and data sources) is expanding fast. The number of non-human entities connecting to remote MCP servers rose 490%, and MCP-related events grew 600%. The report's implication is blunt: the question is no longer only which AI applications run in the environment, but which users and systems those agents interact with and what they can access.
Are attackers exploiting the trend?
Yes, and with measurable intensity. Interactions with malicious AI links — harmful links returned by an AI application that a user clicks or an agent follows — began rising in July 2025, reached 150 per 100,000 users in October, then climbed again from February 2026 to 200 per 100,000 users in March.
"These links are similar to malicious links planted in search engine results, where users may click them believing they lead to legitimate websites," the report states. Attackers promote them through SEO techniques, paid advertisements, or compromised legitimate infrastructure. Netskope also flags artificial intelligence engine optimization (AIEO), an emerging technique attackers use to influence the content AI models surface, plus advertising beginning to appear inside AI applications.
Which applications do security teams block?
The report names the leaders. "Particular Audience is the most frequently blocked AI application in the manufacturing sector, with 45% of organizations restricting access. Emergent follows at 39%, while DeepSeek is blocked by 36% of organizations." Restrictions reflect concerns about where prompts and uploaded files are stored, which internal resources an app can reach, and whether processing happens outside established controls.
What should data teams do?
Netskope recommends a default-deny posture: allow reputable applications already in use, block the rest, and inspect all HTTP and HTTPS downloads across every stream of web and cloud traffic. Concretely:
- Deploy data loss prevention (DLP) policies to detect source code, regulated data, passwords, keys and encrypted data before it reaches personal app instances or AI applications.
- Apply Remote Browser Isolation (RBI) for higher-risk site categories such as newly registered domains.
- Gain visibility over both managed and unmanaged AI applications and API interactions, with real-time controls on sensitive data exposure and unsafe prompts.
- Inspect links returned by AI applications the same way links in search results and email are inspected, before users or agents follow them.
- Red team private AI deployments to find misconfigurations before production exploitation.
Shadow AI is shrinking but not gone. Personal AI account use fell from 75% to 55%, yet the share of users switching between personal and enterprise accounts rose from 9% to 25%. Netskope's remedy is procedural: accelerate review and approval of new AI applications, keep instance-level controls, and maintain a simple request process — because shadow AI rarely disappears entirely.
Google News: Data Governance (Source)
More from Marcus Bennett
Show full bio
News editor covering people, launches and day-to-day practice in Analytics Digest is a professional daily at Analytics Digest.
10 articles