Honeywell Report Reveals Critical Visibility Gaps in OT Security

A significant disconnect exists between how industrial organisations perceive their operational technology (OT) security maturity and their actual ability to track assets, according to a new report from Honeywell.

The 2026 OT Cybersecurity Benchmark Report reveals that while 88% of surveyed security leaders describe their OT programmes as mature or design-led, only 21% maintain a complete inventory of their OT assets.

Visibility and Monitoring Deficits

The survey of 603 leaders across critical infrastructure sectors—including energy, manufacturing, and healthcare—highlighted widespread visibility gaps. Only 33% of respondents reported that OT is fully integrated into a centralised security operations centre (SOC). Furthermore, just 20% of organisations are continuously monitoring more than three-quarters of their connected Internet of Things (IoT) devices.

These visibility shortcomings carry heavy operational costs. Organisations that experienced significant OT cybersecurity incidents reported an average of 16.2 hours of downtime. Among those affected, 21% estimated downtime costs exceeded $100,000 per hour, while 4% reported losses above $500,000 per hour.

Sector-Specific Risk Profiles

Incident rates varied significantly across different industries. The energy and utilities sector reported the highest frequency, with 91% of respondents experiencing a significant OT cybersecurity incident in the past 12 months. Maritime respondents followed closely at 87%. In contrast, only 54% of those in the oil and gas sector reported such incidents.

The healthcare sector also faces unique integration challenges. Only 19% of healthcare respondents stated that their facility and building systems are fully integrated into cybersecurity monitoring and protection protocols.

The Role of AI in OT Security

Artificial intelligence (AI) is expected to become a staple in industrial security, with 99% of leaders anticipating it will impact OT security operations within the next two to three years. Current adoption is already underway, with 72% using AI for threat detection and 68% for continuous monitoring.

However, true autonomy remains limited. Only 23% of organisations currently utilise autonomous or agentic AI for threat detection, indicating that most existing deployments are designed to assist human analysts rather than operate independently.

Honeywell noted that as AI transitions from an assistive tool to an active participant in decision-making, organisations must establish clear decision rights and human oversight to prevent incorrect responses from impacting equipment safety or operational uptime.