Veeam Research: Who is Responsible for Rogue AI Behaviour?

When AI took the world by storm, businesses had a choice, adopt or be left behind. 

But the goal to incorporate this technology into everyday workflows was easier said than crossing, what appears to be, a wide gulf between AI ambition and organisational readiness. 

Research unveiled by Veeam Software at VeeamON London reveals that even as enterprises are embracing AI at pace, many are struggling to build the foundations needed to use it safely and effectively.

The Data and AI Trust Gap report, based on a global survey of 600 senior executives across financial services, healthcare, manufacturing, retail and technology, found that 88% of organisations are already using or piloting AI agents. 

However, only 7% qualify as truly AI-ready. This means incorporating all three of the AI pillars – ambition, visibility and governance. 

More strikingly, 95% of respondents say that data-related challenges have already slowed the progress of their AI wheel.

“Most organisations don’t have an AI adoption problem,” says Anand Eswaran, CEO of Veeam.

He continues: “They have an AI trust problem. The first phase of AI was defined by infrastructure investment, experimentation and acceleration. 

“The next phase will be defined by trust. With the widespread adoption of autonomous AI agents operating at machine speed, the question transitions from whether you can use AI, to whether you can ensure all your data is secure, governed, compliant and resilient. 

“And should something go wrong, can you recover with precision? That’s how you accelerate safe AI at scale without accelerating reputational and operational risk.”

AI’s leadership misalignment problem 

The report points a finger at a glaring disconnect between executive confidence in the technology and its operational reality. 

Business leaders are throwing their weight behind the transformative potential of AI while technical teams are painting a starkly different picture of organisational preparedness.

65% of CEOs believe their organisation maintains a complete inventory of AI systems, compared with just 48% of technical leaders. 

Similarly, more than half (52%) of CEOs believe they actively lead on data strategy, yet fewer CIOs and CISOs share that view at 41% and 38% respectively. 

This disconnected view of the organisational AI war chest raises concerns on AI governance which relies on accurate visibility and clear accountability. 

Who is responsible if AI messes up?

Veeam’s research reveals that this fragmented ownership and misaligned operating disciplines continue to undermine progress, with responsibilities for data, AI and governance often spread across multiple teams.

“When ‘everyone owns it,’ no one can decisively set policy, enforce controls or prove outcomes,” the company highlights. 

Numbers back this up, showing that organisations that have a clearly defined ownership – where CISOs own agentic AI risks – are 24% more likely to detect “rogue AI behaviour”.

Conversely, organisations with shared ownership are 47% less likely to detect them.  

The hidden problem of shadow AI 

Shadow AI is a growing enterprise risk that is now mainstream, with 95% of organisations reporting unauthorised AI usage within their workforce.

Though 93% view shadow AI as a significant risk, only 25% provide employees with approved alternatives.

Concerningly, research shows that only a minority among the organisations already deploying AI could quickly identify what data an AI system accessed, what actions it took or which decisions it influenced.

Just 40% of leaders expressed strong confidence in their ability to isolate and reverse an AI-related failure with precision.

This at a time when regulatory scrutiny is tightening its grip. More than six in ten organisations say the EU AI Act has already influenced their investment decisions over the past year, while maintaining audit trails for AI decisions has emerged as a major compliance concern among 47%. 

Building trust in AI

The research suggests that organisations achieving the greatest success with AI are not necessarily those moving the fastest, but those building strong foundations first. 

The study also found that nearly half of CEOs believe trusted and compliant data could unlock more than 25% revenue growth. Yet many organisations admit their data still needs to be more accurate, accessible and up to date before those benefits can be realised.

“The findings here leave no room for doubt. When 95% of executives say data challenges are already slowing their AI progress, the bottleneck isn’t the model – it’s trusted, governed, recoverable data,” adds Anand. 

“Veeam is building the Data and AI Trust layer to give enterprises the visibility, control and precision recovery needed to scale AI safely and deliver real business value.”