AIToday

Enterprise AI agents outpaced safety controls by design

VentureBeat AI4h ago

Key takeaway

Enterprises knowingly deployed AI agents faster than they built controls to manage them, according to VentureBeat Research surveys conducted in June across five governance layers. Now they are retrofitting: 57 to 68% plan to switch or add vendors within 12 months, and roughly a third aim to move within the quarter. The five measured controls are identity (authorization), evaluation (quality), cost telemetry (expense tracking), context layer (data supply), and orchestration.

Summaries like this, in your inbox every morning.

Sign up free →

3 Key Points

  • What happened

    VentureBeat Research surveyed enterprises across five control layers for AI agents in June and found that 57 to 68% plan to switch or add vendors within 12 months to retrofit governance. Roughly a third plan to make these moves within the quarter. The five control layers measured are identity (which agent can do what), evaluation (whether work is good), cost telemetry (tracking agent costs), context layer (supplying business data), and orchestration.

  • Why it matters

    Enterprises deployed AI agents before building the controls needed to manage them — and they did this deliberately, according to the survey. They are now budgeting and moving quickly to catch up with their own governance standards. This suggests enterprises recognize both the risk and the opportunity cost of uncontrolled agent deployments.

  • What to watch

    The speed of remediation: roughly a third of enterprises plan vendor changes within the quarter, indicating governance gaps are seen as urgent. How vendor consolidation and new entrants reshape the agent-control market may signal which governance approaches enterprises are betting on.

In Depth

VentureBeat Research conducted five parallel surveys in June to measure enterprise readiness across the agentic AI stack. The core finding was striking in its directness: enterprises deployed AI agents ahead of the controls needed to manage them — and they did this knowingly. This was not a gap born of ignorance but of deliberate prioritization.

The research identified five critical control layers that enterprises must build before they can trust an agent in production. Identity governs which agent is allowed to do what, under whose credentials — the foundational authorization layer. Evaluation determines whether the agent's work is any good, addressing quality assurance. Cost telemetry tracks what each agent costs to run, surfacing the expense of inference at scale. The context layer supplies the business data and definitions agents need to operate. Orchestration coordinates agent workflows across the enterprise.

The governance deficit is real, but so is the remediation effort. In each of the five control layers, 57 to 68% of enterprises plan to switch vendors or add new ones within the next 12 months. Roughly a third, depending on the layer, plan to move within the quarter. This rapid pace signals both urgency and active budgeting: enterprises are not waiting for a single platform to mature but are actively replacing and layering new tools to close the governance gap. The variation by layer suggests that some controls (perhaps identity or cost telemetry) may be more mature in the vendor landscape than others, driving different timelines for different enterprises.

Context & Analysis

The VentureBeat Research surveys, fielded in June across every layer of the agentic stack, reveal a deliberate trade-off: enterprises prioritized deploying AI agents over building the governance infrastructure to manage them safely. This was not a surprise or an oversight but a calculated move, likely driven by competitive pressure to demonstrate agent capabilities and business value before hardening controls.

The immediate consequence is a retrofit cycle. Enterprises are now budgeting significantly to add or replace vendors in identity, evaluation, cost telemetry, context, and orchestration — the five foundational control layers. The fact that 57 to 68% across all five layers plan vendor changes within 12 months, with roughly a third targeting the quarter, suggests two things: first, governance gaps are recognized as urgent, not theoretical; second, no single incumbent vendor has established dominance in these control layers, creating a fragmented and competitive market where enterprises are still shopping for the right mix of tools.

FAQ

What are the five control layers VentureBeat measured?
Identity (which agent is allowed to do what, under whose credentials), evaluation (whether the agent's work is any good), cost telemetry (tracking what each agent costs to run), the context layer (supplying business data and definitions to agents), and orchestration.
How quickly are enterprises planning to retrofit?
In each of the five control layers, 57 to 68% of enterprises plan to switch vendors or add new ones within 12 months, and roughly a third plan to move within the quarter, depending on the layer.
Did enterprises deploy agents knowingly without controls?
Yes. The central finding of VentureBeat Research's five parallel surveys is that enterprises deployed AI agents ahead of the controls needed to manage them — and they did it knowingly.

Get the latest Large Language Models news every morning

AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.

Free · takes 30 seconds · unsubscribe anytime

Discussion

No comments yet. Be the first to share your thoughts!

Log in to join the discussion

Related Articles

Stay ahead with AI news

Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.

Get Started Free

Free · takes 30 seconds · unsubscribe anytime