Routine employee activity across corporate systems carries an average annual cost of $19.5 million per organization. That figure comes from the 2026 Cost of Insider Risks Global Report, conducted by the Ponemon Institute and based on data from 354 organizations that experienced one or more material insider related incidents over the past year.
Negligent or mistaken insiders account for the largest share of financial impact. These incidents generate $10.3 million in annualized cost per organization, with an average cost of $747,107 per incident and 13.8 incidents per organization each year. Malicious insiders account for $4.7 million in annualized cost, and credential theft incidents add another $4.5 million.
Containment remains the most expensive phase of the incident lifecycle. The average cost of containment is $247,587 per incident. Organizations spend an average of 67 days containing an insider event. Incidents resolved in under 30 days carry annualized costs of $14.2 million. When containment extends beyond 90 days, annualized costs reach $21.9 million.
Negligence and AI usage intersect
Employee negligence continues to dominate incident frequency, and its financial impact has increased year over year. The annualized cost tied to negligent or mistaken insiders rose from $8.8 million in 2024 to $10.3 million in 2025.
GenAI has altered how employees access and share information. Most organizations report changes in workforce behavior tied to AI use, yet only a small share have formally embedded generative AI into business strategy. Concern about unauthorized AI creating unseen data exfiltration paths remains widespread, and only a limited portion of organizations have integrated AI governance into insider risk management programs.
Investigations identify recurring patterns tied to AI tools. Employees input internal documents, legal materials, source code, architecture diagrams, and strategic plans into public AI platforms such as ChatGPT, Claude, Gemini, Perplexity, and Grok. AI meeting assistants generate transcripts and summaries that include sensitive internal discussions and personal data. AI browsers and agents access enterprise systems and execute tasks that bypass traditional logging controls.
“To mitigate shadow AI risks without stifling productivity, organizations must shift from a block first approach to an audit and educate model,” Raj Koo, CTO of DTEX, told Help Net Security. “Technologies that understand intent, inspect prompts, and analyze usage patterns are critical. Instead of outright blocking, real time nudges at the point of risk can guide employees toward safer practices.”
Koo tied that model to measurable outcomes reflected in the benchmark data.
“This approach reduces containment time for insider incidents by 17 percent, from 81 to 67 days, and prevents at least seven major incidents annually, saving $8.2 million in breach costs,” he said. “By embracing visibility and education, organizations can enable the workforce while safeguarding sensitive data.”
Visibility into AI agent activity remains limited across many organizations, increasing uncertainty around how these tools interact with enterprise data.
Investment and measurable savings
Dedicated insider risk management programs correspond with fewer incidents and lower financial impact. Sixty three percent of organizations operate a formal insider risk management program. Organizations with a program avoid an average of seven insider incidents each year, resulting in roughly $8.2 million in avoided breach costs.
Spending on insider risk management has increased over the past two years. In 2025, organizations allocate an average of 19 percent of IT security budgets to insider risk management, up from 8.2 percent in 2023. Many organizations increased insider risk budgets in the past year, and a majority expect further increases in 2026.
Identity management and behavioral monitoring technologies deliver the largest cost reductions. Privileged access management produces average savings of $6.1 million. User behavior analytics generates $5.1 million in savings. Forty two percent of organizations use AI to detect or prevent insider risks, and reduced false positives rank among the most frequently cited benefits.
Containment times have declined alongside increased investment. Average time to contain an insider incident fell from 86 days in 2023 to 67 days in 2025.
AI agents extend the insider boundary
AI agents operate inside enterprise workflows, either across business units or within security teams. Many organizations view agentic AI as important for early insider risk detection, and concern about malicious use increasing data theft risk is common. Even so, only a small share classify AI agents as equivalent to human insiders, leaving gaps in how these systems are governed and monitored.
Koo said insider risk programs need to treat AI agents as operational identities with defined access controls.
“To classify the risks posed by AI agents, organizations must view them as dynamic, non human identities governed by functional access controls,” he said. “The core of this approach is the lethal trifecta framework, which highlights the highest risk agents, those with overlapping capabilities in three areas: access to sensitive data, ability to communicate externally, and exposure to untrusted content.”
He said effective oversight depends on tighter privilege controls and better logging.
“Mitigating these risks requires tiered access controls that differentiate between read only and write access permissions,” Koo said. “Reasoning logs close the logging gap by capturing the intent behind autonomous actions. Micro privileges and human in the loop approvals for high stakes decisions are essential guardrails.”
Insider risk costs vary across sectors. Health and pharmaceutical organizations report average annualized costs of $28.8 million. Technology and software organizations report $24.2 million. North American organizations report $24.0 million, above the global average.
Download: Picus Security’s Red Report 2026
