85% of IT leaders agree that secure identity and access management is critical for scaling AI safely.

Meeting intelligence tools Otter.ai and Read.ai are present in 74.2% and 62.5% of organizations respectively; Gamma is present in 52.8% of organizations, Cursor in 48.4%, and ElevenLabs in 45.2%.

76% of defenders say AI agents or AI assistants now handle more than 10% of their workload.

74% of leaders remain concerned about security risks from AI.

In MCP registries, for every server provided by a verified technology vendor there are up to 15 lookalike servers from untrusted sources.

At least 700 million people use leading AI systems weekly.

28% of organizations can reliably trace agent actions to a human or system across all environments.

58% of organizations estimate they currently have between 1 and 100 agents deployed.

Automated verification of infrastructure security surged by more than 50%.

44% of organizations use or plan to use static API keys and 43% use or plan to use username/password combinations for agents.

18% of IT and security professionals are highly confident their current IAM systems can manage agent identities effectively, while 35% report moderate confidence, 29% report slight confidence, and 18% report no or uncertain confidence.

Among retail Pindrop customers, non-live fraud increased 56% month-over-month in November 2025.

Teams using attack intelligence to track emerging AI vulnerabilities increased by 10%.

Nearly one-third of all healthcare email incidents were attributed to vendor and business associate email exposure, making it the most frequent attack pattern.

AI fraud surged 1210% in 2025.

Application of custom rules to automated code review tools to catch issues unique to AI-generated code increased by 10%.

In Q4 2025, callback phishing increased from 3% to 18% of all phishing incidents, a 500% spike.

In Q4 2025, Business Email Compromise accounted for 51% of all email fraud cases.

Diversion tactics (fraudulent invoices, fake payroll requests) accounted for 18% of BEC incidents in Q4 2025.

14.4% of AI agent configuration files grant arbitrary code execution permissions for Node.js.