Only 32% of organisations conduct daily API risk assessments.

Nearly half of employees (46%) admit to pasting company information into public AI tools.

37% of organisations have a dedicated API security solution.

64% of US consumers have interacted with AI chatbots more frequently in the past year.

44% of senior IT and infrastructure leaders in the business and professional services industry are planning to adopt Secure Access Service Edge (SASE) solutions in the next 12 months

91% of organisations have taken new steps to protect operational resilience due to waning federal support.

48% of senior IT and infrastructure leaders in the business and professional services industry identified simplifying operations and reducing IT burden as their top strategic networking and security priority.

61% of cybersecurity and IT professionals intend to adopt AI and machine learning (ML) for security within the next 12 months. This adoption rate outpaces general manufacturing by 12 percentage points.

Over half (53%) of organisations using agentic AI say they are already deploying it, or plan to, for customer-facing roles.

61% of cybersecurity professionals plan AI adoption as manufacturing faces increasing cyber risks.

Among external risks to manufacturing, cybersecurity is ranked highly at 30%, coming in second only to inflation and economic growth, which stands at 34%.

One in 10 employees describe AI adoption in their workplace as "the Wild West," meaning it's unregulated, unclear, and out of control.

29% of U.K. IT and security decision-makers have delayed or cancelled contracts due to growing U.S. cybersecurity instability.

Almost a fifth (18%) of organisations host between 501–1000 AI agents.

37% of US consumers trust interactions over the phone.

15% of employees who hide their AI use worry it will be seen as risky.

57% of UK organisations recovered from backups when hit by ransomware.

59% of UK organisations have immutable backups.

54% of US consumers trust interactions in person.

62% of US consumers believe AI agents are easier to trick than humans