Survey of 100 Security Leaders Finds Trust and Data Access Barriers to AI Agent Adoption

· Original Report (PDF) ↗

Summary

Strike48’s survey of 100 security leaders found that trust concerns and limited log access are barriers to AI agent adoption in security operations; 36% reported agents in production for at least one use case.

Key points

  • The survey says 84% of respondents believe AI agents should handle SOC L1 work, but only 22% are ready to fully automate those tasks.
  • Only 36% reported having agents in production for at least one use case; 34% said they were running proofs of concept.
  • Trust concerns were widespread: 64% selected three or more, and 52% cited AI trust as a top adoption barrier.
  • The report says 84% cannot access all their log data for investigations, while 65% had an investigation stall or be limited because data was inaccessible.
  • Respondents identified hot-storage costs as a data-access concern: 80% said the cost was painful or a major budget issue.
  • The report recommends improving agent auditability and human oversight, expanding data access, and introducing agents alongside existing security tools.
  • Strike48 produced the survey and promotes its own agentic security platform in the report.

Article Details

Publisher
Strike48
Report Period
2026
Scope
Survey of cybersecurity leaders at enterprise and mid-market organizations on agentic security adoption, trust, and data visibility.
Sample Size
100 security leaders
Key Statistics
  • 84% said their current tools cannot access all their log data for investigations.
  • 65% reported at least one investigation stalled or was limited by inaccessible data.
  • 36% had agents in production for at least one use case.
  • 22% were ready to fully automate L1 work.
  • 64% selected three or more trust concerns.
Recommendations
  • Connect AI agents to the existing security stack rather than replacing it.
  • Ensure agents can access data across multiple sources, including log data in S3 and observability tools.
  • Train L1 staff to manage agents and use adjustable human-in-the-loop controls as agent autonomy increases.

Vendors

Industries