Summary
- Perforce surveyed more than 600 practitioners on AI, real-time workflows, infrastructure, and digital-asset management.
- Half of respondents cited job insecurity as an AI concern, while content quality and compliance were close behind.
- Productivity gains are increasing the importance of version history, provenance, review, and governance across AI-assisted workflows.
Generative AI is accelerating production workflows across software, media, entertainment, automotive, and manufacturing, but the same adoption is increasing pressure on organisations to prove where digital assets came from, how they changed, and whether outputs can be trusted.
Perforce surveyed more than 600 practitioners for its 2026 State of Real-Time Workflows report, produced with Amazon Web Services. The findings point to measurable productivity gains in sectors that have moved furthest with generative AI while also exposing concerns over compliance, content quality, ethics, and job security.
Job insecurity was the most common AI-related concern globally, cited by 50% of respondents. Content quality followed at 49%, compliance at 48%, and reduced creativity at 36%.
Productivity gains were strongest in media and entertainment and in automotive and manufacturing. Perforce said 48% of media and entertainment respondents and 41% of automotive and manufacturing respondents reported productivity improvements of between 11% and 50% after adopting AI.
The regional picture was uneven. Respondents in Asia-Pacific reported the strongest AI-driven acceleration, while concern over AI-related job losses was highest in Latin America and remained substantial in North America.
Perforce’s survey also found that version-control adoption had reached 94%, up from 86% in 2025, while hybrid cloud and on-premises server configurations rose to 27% of respondents from 10% a year earlier. Those figures suggest that faster content creation is arriving alongside more complex infrastructure and a larger volume of digital artefacts to govern.
Provenance becomes part of operational control
The governance challenge is not simply whether an organisation uses AI. It is whether teams can trace AI-assisted changes through existing development and production systems with enough clarity to satisfy quality, compliance, intellectual-property, and audit requirements.
That is harder when AI systems can generate code, 3D assets, scripts, media, documentation, and other production material at a much higher rate than manual workflows. More output means more versions, more dependencies, and more decisions that need to be associated with a responsible person, process, or system.
Visibility into AI-generated code and code changes is particularly important where software feeds into regulated, safety-critical, or revenue-generating products. A productivity gain can be offset quickly if teams cannot reconstruct how a change entered a repository, which model or tool produced it, what review it received, and whether the final output complied with policy.
The same issue applies outside conventional software development. Real-time 3D engines and digital-creation platforms are now embedded in film, television, automotive, manufacturing, and engineering workflows, which means provenance and version history increasingly span mixed technical and creative assets rather than source code alone.
Perforce’s findings also show why AI governance is becoming closely connected to workflow architecture. Organisations that already have disciplined versioning, review, approval, and access controls have a clearer place to insert AI-generated work than those relying on informal file sharing or weakly governed production pipelines.
The survey does not establish that AI adoption is making workflows less secure or less compliant in every organisation. It shows instead that the productivity benefits are arriving faster than confidence in the surrounding controls.
That gap is likely to become more visible as AI use expands from individual assistants into automated and agentic workflows. The more decisions software makes on behalf of developers, designers, and engineers, the more important it becomes to retain a reliable record of what changed and why.
Perforce’s data therefore points to a practical limit on AI-driven acceleration: teams can produce more, but the value of that output depends increasingly on whether the organisation can establish provenance, enforce policy, and recover a trustworthy history of the work.
