50.0% of large firms say AI is used company-wide

The survey lands at a point where AI tools have moved from experiments into day-to-day development. Among employees at large firms, half now report company-wide AI use, and their non-engineer colleagues have joined in: 65.0% of large-company respondents said their company's non-engineers are using AI tools for development, against 40.0% overall. That spread is exactly why the survey's answers about checking AI output carry weight — code now reaches the whole company, but the survey indicates review has not kept pace. Only 9.1% of respondents at firms with 5,001 or more employees picked company-wide AI security and governance rules as a needed measure, far below the 38.2% at firms with 101 to 1,000 employees, even though the larger firms report the deeper governance gaps in this survey.

Where engineers sit shapes the answers. Among IT department respondents, 64.3% wanted cost-visibility and management tooling, well above the overall figure, and 41.4% chose it as a second priority, 15.4 points ahead of non-IT respondents. Non-IT respondents instead leaned toward ready-made standards for AI code development at 24.7%, versus 13.8% for IT staff. The free-comment section points to why: one respondent said AI was introduced with the instruction that final checks stay with humans, and another described a system that became an error after an unexpected event exceeded what AI could handle.

The open question is whether large firms turn these findings into internal rules. The survey suggests the pressure is real, since 55.0% of those using AI in development said AI-related defects and issues exceed what IT can handle. What remains unclear is whether the safeguards arrive through the code-checking tools most respondents asked for, or through the governance guidelines that smaller firms favored more. For IT staff at large companies, the practical test will be whether they can combine usage and cost visibility with a repeatable way to check AI-generated code.