New Research The AI Fog Report: When the Signal Becomes the Noise
Explore how AI is reshaping complexity, alert noise, and trust in modern security operations, from the people doing the work.
Smarter signals, stronger decisions
Based on an independent survey of 595 cybersecurity decision-makers and practitioners across the United States and the United Kingdom.
53.4%
say their security tools add complexity or only mixed value
86.1%
report increased alert volumes in the past 12 months
24.5%
fully trust AI to remediate threats without human oversight
The AI fog challenge
Three interconnected trends are shaping how security operations work now.
Complexity
AI is entering security environments that are already complex, adding new data, workflows and governance requirements alongside new opportunities for efficiency.
Context
Security teams increasingly face a context challenge, not just a volume problem. Duplicate alerts and AI-generated outputs without enough context make it harder to determine what matters.
Trust
Adoption is accelerating faster than operational trust. Explainability, human override, and a proven low false-positive track record are essential for building the confidence teams need to grant AI greater autonomy.
From more AI to better action
The report points to five priorities for turning more security data into faster, more confident action.
Context-aware prioritization
Prioritize alerts based on the organization or client environment, rather than ranking them in isolation.
Unified visibility
Reduce duplicate alerts across tools, tenants, and telemetry sources.
Faster detection-to-containment
Use AI to accelerate response while reducing false positives and unnecessary human review.
Trust built incrementally
Build confidence through explainability, human override, and scope-limited autonomy.
Operating models that fit reality
Design security operations for MSP scale, SMB capacity constraints, and midmarket integration complexity.
Better action starts with better signal
Better security outcomes aren’t driven by more detections alone. They depend on reducing duplication, adding environment-specific context, supporting faster decisions, and building the trust to act with confidence.
