Executives warn AI expansion is outrunning controls

May 15, 2026 3 min read

Veeam said companies are adopting artificial intelligence faster than they are building governance and audit controls, while nearly half of executives admitted their confidence in scaling AI is based more on intuition than auditable evidence.

Table of Contents

The adoption of artificial intelligence is advancing faster than companies’ ability to govern it and demonstrate that their systems operate securely and remain auditable.

The rapid expansion of artificial intelligence within enterprises is beginning to expose a growing gap between technology adoption and organizations’ ability to govern, control and audit AI-driven systems.

Veeam Software announced the launch of its Data Trust and AI Maturity Model, a framework designed to help organizations assess how prepared they are to manage artificial intelligence in enterprise environments where autonomous agents already operate on critical data at increasing speed and scale.

The company said many organizations have already moved beyond the initial stage of AI adoption but still face difficulties deploying it operationally in a secure, consistent and auditable way for regulators, boards of directors and compliance environments.

A study conducted by Emerald Research Group on behalf of Veeam found that companies have advanced faster in implementing AI than in building controls related to identity, governance, resilience and data management.

Related Posts:

According to the research, the challenge is no longer simply using artificial intelligence, but understanding, validating and controlling the decisions these systems are beginning to make autonomously.

According to Anand Eswaran, CEO of Veeam Software:

“The Data Trust and AI Maturity Model gives leaders a clear and objective way to understand their real position, identify execution gaps and prioritize the capabilities needed to operationalize trust in AI.”

The company said the model evaluates organizational maturity across 12 dimensions and measures progress through five stages, from ad hoc levels to advanced leadership stages.

The goal is to identify where controls exist and where they break down under real-world conditions, and what should be prioritized to strengthen governance, trust and resilience in AI-driven environments.

A study reveals a growing gap in trust and preparedness.

The model is based on research conducted with 300 senior executives responsible for technology, security, data and enterprise risk.

Key findings from the study include:

AI is no longer experimental. Nearly seven in ten companies report that artificial intelligence is already integrated into multiple business functions or is core to operations, meaning AI systems and agents interact daily with sensitive data, customer records and decision-making processes.

Executive confidence is high. Eighty percent of leaders say they are confident in their ability to scale AI securely over the next two years.

Confidence often lacks evidence. Nearly half of executives acknowledge that their confidence is based more on intuition than on auditable evidence.

As AI expands, execution challenges are growing. Fifty-two percent of organizations reduced AI initiatives over the past 18 months, four in ten experienced delays and 28% paused projects entirely.

Barriers to progress are primarily operational. These include shortages of AI and machine learning expertise (43%), integration challenges with existing systems (33%), regulatory uncertainty (25%), data quality limitations (20%) and concerns related to explainability (19%).

Governance continues to lag AI deployment. Although nearly nine in ten companies say they have formal AI policies, only one in three says it can immediately produce complete audit evidence.

Taken together, the findings show that artificial intelligence deployment is advancing faster than the operational maturity required to sustain it in critical environments.

From adoption to demonstrable trust.

Veeam said the model aims to move the conversation from simple implementation toward the ability to demonstrate operational trust under real-world conditions.

To achieve this, the company structured the framework around four pillars:

Understanding: visibility and context around data, AI assets and risks.

Security: identity controls, privacy and data protection.

Resilience: backup, recovery and operational continuity for AI-driven environments.

Enablement: preparation of trusted data to support trusted AI adoption.

Krista Case, principal analyst at theCUBE Research, said:

“AI success depends on the strength of the data foundation, but that is precisely where companies remain exposed.”

The analyst added that although many organizations already have mature AI deployments, fewer than one-third back up at least half of the data generated by these systems, increasing risks associated with inference, corruption, poisoning, exfiltration and attacks targeting data layers.

Veeam said the model will be implemented through a Data Trust and AI Maturity Assessment, a consulting service that will provide comparative profiles, prioritized recommendations, improvement roadmaps and tracking tools for audits and boards of directors.

The company officially introduced the initiative during VeeamON 2026 New York and said the assessment will become globally available later this year.