Netskope reports reveals rising enterprise AI data leakage

As financial services in India push deeper into AI-led transformation, a new report from Netskope Threat Labs points to a growing concern, Enterprise AI data leakage is quietly becoming a real risk. The findings show a clear tension. On one side, organisations are adopting AI for speed and efficiency. On the other, sensitive financial data is increasingly exposed through everyday workflows. This shift is subtle, but its impact is significant.

Where the real risks are building up

The report highlights that 59% of all data policy violations linked to generative AI involve regulated financial data. That’s not a small number—it shows that the most sensitive data is also the most exposed.

AI adoption is already widespread. Around 70% of users actively use GenAI tools, while 97% interact with AI-powered applications in some way. The challenge is deeper than usage. Nearly 94% of these applications depend on user data for training, increasing the chances of unintended exposure.

The overlooked problem: everyday user behaviour

One of the more subtle risks comes from user habits. Even as companies move towards safer, enterprise-managed tools, gaps remain. Enterprise GenAI usage has increased from 33% to 79%, personal app usage is declining, but not disappearing, and 15% of users still switch between personal and work accounts. This “account switching data leakage” creates blind spots. Sensitive financial data can easily move between controlled and uncontrolled environments without clear visibility.

Beyond AI: familiar platforms, new risks

The report also points to risks outside GenAI. Personal cloud applications continue to be a weak link, with 65% of policy violations involving regulated data.

Common workplace tools are part of the story: LinkedIn accounts for 92% of such violations, Google Drive follows at 84%, and GitHub stands out as the top malware delivery platform, affecting 11% of organisations

These numbers show that AI security risks in financial services are not limited to AI tools alone. They extend across the broader digital ecosystem employees use every day.

Why control, not restriction, is the real shift

Organisations are not stepping back from AI, they are trying to use it more safely. The move towards enterprise-managed platforms shows intent, but it also highlights the need for stronger controls.

The report suggests a layered approach: Monitoring all web and cloud traffic, Blocking unnecessary applications, Using DLP for GenAI financial data protection, and Enabling secure access through controlled environments

This is less about limiting AI and more about building systems that can handle its risks without slowing innovation.

Final takeaway

The Netskope report makes one thing clear, Enterprise AI data leakage is not a future problem; it’s already happening.

For financial institutions, the challenge is not whether to adopt AI, but how to do it without losing control over sensitive data. The answer lies in better visibility, smarter controls, and a deeper understanding of how people actually use these tools every day.