New research from AvePoint indicates that enterprises are deploying AI agents faster than they are establishing corresponding governance frameworks. This gap is associated with reported security incidents, deployment delays, and operational risks across digital environments.
According to the State of AI Report, visibility challenges related to AI usage have increased, with 17.6% of organisations reporting they are unable to track unsanctioned AI tool use, up from 6.3% previously. For AI agents specifically, 21.1% of organisations report they are unable to manage unauthorised agent activity.
Adoption of AI agents in workplace operations is widespread, with 46.9% of employees reported to use AI agents regularly. At the same time, 88.4% of organisations reported experiencing agent-related security breaches over the past year, indicating a gap between deployment and control measures.
There is also a contrast between reported confidence and reported outcomes: 82.7% of respondents believe they are able to prevent unauthorised AI data access, although some within this group also report having experienced unauthorised access incidents. This suggests a discrepancy between perceived and actual control in some organisations.
For managed service providers (MSPs) and cloud providers, this situation is associated with increased demand for services related to data governance, AI agent oversight, and deployment readiness.
Governance challenges are also reported to affect implementation timelines. Approximately 90% of organisations report delays averaging around six months in AI deployments, citing security and data management concerns. This can extend pre-deployment planning and increase the need for validation and oversight processes.
The growth of AI-generated content is also relevant to these challenges. AI is currently estimated to produce 35.5% of enterprise data, with projections suggesting this could rise to 42.1% within a year. Many organisations also manage large volumes of legacy data, which contributes to the need for ongoing data classification and modernisation.
In response, organisations are increasing investment in governance and monitoring tools, including categories such as AI Agent Management Platforms (AMPs) referenced in Gartner research. These tools are being used to support visibility, policy enforcement, and operational control.
Chris Shaw, Channel Director for UK and Ireland at AvePoint, notes that the industry’s focus on AI capability has overshadowed fundamental governance weaknesses. “These days, everyone talks about how powerful AI has become… but this discourse has largely distracted from the fact that many organisations face basic data protection and governance issues that are exposing them to serious risks,” he said. “Almost 9 in 10 organisations have experienced an AI‑related breach in the last year… that really shows you the scope of the problem.”
Overall, the findings suggest that effective AI outcomes are closely linked to governance maturity, including visibility, control mechanisms, and data management practices.
