The people responsible for deploying AI inside large organizations are sending a clear signal: the technology is moving faster than their ability to control it.

At the Gartner IT Symposium/Xpo in Gold Coast, Australia this week, the firm’s research delivered a frank assessment to more than 1,500 CIOs gathered for what Gartner describes as the world’s most important annual gathering of IT executives. The headline number is striking: 86% of CIOs believe that AI risk is growing faster than AI value.

That is not a fringe view. It is the majority position among the people actually running enterprise technology decisions.

What Gartner Actually Said

Gartner Distinguished VP Analyst Daryl Plummer put it plainly from the opening session: “Trust in these vendors is not warranted yet. They are not enterprise grade. They don’t understand enterprise terms and conditions. They don’t understand enterprise liability.”

This is a significant statement from one of the world’s most influential technology research firms. Plummer was not talking about small AI startups. He was talking about the leading frontier AI labs that businesses are currently rushing to adopt.

The core problem, according to Gartner, is structural. Foundation models are being treated as enterprise software products, but they do not behave like enterprise software products. They have frequent model updates with no legacy support commitments, evolving safety policies, unclear liability frameworks, and no consistent audit trails. These are not minor gaps. They are fundamental to how enterprises buy, deploy, and govern technology.

The result is a trust deficit that no amount of benchmark performance can fix.

The “Careless Consumption” Problem

Beyond the vendor readiness issue, Gartner’s research surfaces a second problem that is entirely internal to organizations: workers using AI frivolously, incorrectly, or without appropriate oversight.

Gartner calls this “careless consumption.” It emerges when early AI successes create more internal demand than IT departments can safely manage. When people see colleagues use AI to write emails faster, they adopt it everywhere, including situations where it generates poor outputs that still end up in business processes.

The data is specific: 40% of workers have encountered AI slop (low-quality, AI-generated content that appears usable but contains errors, hallucinations, or misleading information). Gartner’s analysis found that deciphering such content typically takes around two hours. Across a 1,000-person organization, that adds up to roughly $9 million in lost productive work per year.

That number should get the attention of every business owner or CFO who has been measuring AI ROI only on the output side. The input cost of cleaning up AI mistakes is invisible in most productivity frameworks.

The Scaling Problem

A separate Gartner survey, conducted between January and April 2026 across a broad set of organizations, found that only 22% have successfully scaled AI across multiple business units or adopted a genuinely AI-first approach. This is after years of pilots, investments, and internal campaigns.

The gap between “experimenting with AI” and “running AI at scale with measurable returns” remains wide. Eighty-five percent of functional leaders plan to increase AI spending in 2026, but spending more on a capability that lacks proper governance and trained users does not automatically produce better outcomes.

What This Means for Business

This research reflects what many business owners and operations leaders are experiencing in practice, even if they have not put numbers to it yet.

The AI vendor ecosystem is genuinely powerful but genuinely immature by enterprise standards. Models change without notice. Outputs vary. Liability sits in murky contractual territory. These are not theoretical risks. They affect real deployments, real workflows, and real business outcomes.

Three things business leaders should take from this Gartner research:

1. Vendor maturity is a real selection criterion. Not all AI providers operate with enterprise-grade reliability, versioning, and liability frameworks. Before scaling a deployment, understand what happens when the model changes, who is responsible for errors, and what audit trail exists for regulated decisions.

2. Governance must precede scale. The “careless consumption” problem is not a technology failure. It is an organizational failure. Companies that have not built policies, training, and oversight structures for AI use are accumulating invisible risk in their workflows. A policy introduced after the fact tends to face more resistance.

3. Data foundations determine outcome quality. The organizations in the 22% that have successfully scaled AI share a common characteristic: they invested in data infrastructure and data literacy before deploying AI on top of it. Gartner’s research consistently shows that AI ROI is four times higher in organizations with strong data foundations.

The Bigger Picture

It is worth noting when this finding lands. Gartner delivered this message to CIOs on the same week that Dreamforce 2026 is running in San Francisco, where Salesforce is making the case for an “agentic enterprise.” These are not contradictory signals. They are the two sides of the same reality: the technology is advancing rapidly, and the organizational and vendor infrastructure required to make it safe and effective is catching up.

For Enterprise DNA clients, this is familiar ground. The reason we focus so heavily on data literacy, structured AI implementation, and building internal capability alongside deploying AI services is that the returns from AI are not automatic. They depend on the quality of the foundations underneath.

The 86% of CIOs who see risk outpacing value are not wrong. They are honest about what their organizations actually look like right now. The question is what they do next.

---

Enterprise DNA’s Omni Advisory service helps business leaders design AI strategies that account for both the opportunity and the governance requirements. Book a discovery call to discuss what an enterprise-ready AI approach looks like for your business.