Generative AI: taking back control, SASETY’s vision for companies

OP-ED · 21 MAY 2025

By Jérôme Beaufils, CEO of SASETY · Lire en français

Introduction

Generative AI is rapidly transforming the way people work. Content generation, coding assistance, task automation: the potential is immense. But mass adoption comes with growing risks: loss of control over data, leaks of sensitive information, non-compliant use, shadow IT and obstacles to enforcing security policies.

The GenAI Sprawl phenomenon

The uncontrolled proliferation of generative AI tools, or GenAI Sprawl, is fuelled by massive and often disorganised adoption:

65% of companies already use generative AI tools in at least one business function,
80% of decision-makers have tried these tools, and 20% use them every day,
44% of French users use it in both a personal and a professional context,
And yet 68% of them do not tell their managers, revealing largely hidden use.

With the rise of products such as ChatGPT, Microsoft Copilot, Bard and Claude, or tools built into SaaS applications (M365, Notion, Canva, etc.), many employees are starting to use these technologies on their own, often outside the control of the IT or security team.

This creates dispersion at several levels:

Tools: use of a large number of GenAI applications, often unvalidated or unknown,
Users: varied, inconsistent uses depending on departments, profiles or locations,
Data: potential sending of sensitive information to external platforms without oversight,
Security policies: no way to guarantee compliance and the application of consistent rules.

The risk? Loss of visibility, an explosion of the attack surface and massive exposure to data leaks or processing that breaches the GDPR.

GenAI Sprawl is therefore the AI equivalent of Shadow IT for SaaS applications: a phenomenon that must be brought under control quickly, or innovation risks becoming a major source of vulnerability.

A unified technology response with the Cato Networks SASE platform

Against this backdrop, the 100% cloud-native SASE platform from Cato Networks takes an integrated approach, combining the essential components to govern the use of generative AI in the enterprise.

Two pillars are essential: CASB (Cloud Access Security Broker) and DLP (Data Loss Prevention).

CASB: see, understand, control

Cato's CASB automatically discovers cloud applications, including those from Shadow IT and generative AI usage. It provides:

Complete visibility over the tools in use,
A classification of AI applications by risk,
Granular access policies (read-only, upload ban, role-based segmentation, etc.),
Real-time monitoring of user actions, including via API for services such as Copilot, Entra ID or Box.

DLP: preventing sensitive data leaks

Cato's built-in DLP module monitors all network traffic (SaaS, web, datacentre, cloud) and identifies attempts to transfer critical data. It relies on:

More than 350 preconfigured sensitive data types,
The ability to create custom rules depending on the context,
Block, alert or log actions depending on the AI usage detected.

Modern DLP also makes it possible to identify suspicious prompts, public sharing and indirect leaks of company data through poorly governed AI.

Full control of GenAI usage

Cato's combined CASB + DLP approach secures every step of using generative AI:

Discovery and classification of tools (Copilot, ChatGPT, etc.),
Monitoring of user actions (prompts, uploads, sharing),
Banning or limiting risky uses,
Preventing leaks of sensitive information,
Compliance with internal policies and regulations (GDPR, ISO 27001, NIS2).

SASETY's expert support: a key success factor

To get the most out of these technologies, expert support is essential. SASETY, a strategic partner of Cato Networks, makes their integration easier by ensuring:

Fast, smooth implementation,
Adaptation to the specific needs of each organisation,
Skills transfer to make IT teams autonomous,
Targeted training on analysing GenAI usage,
Continuous monitoring and optimisation of security policies.

As a managed cybersecurity services provider, SASETY also keeps SASE and SSE technologies in operational condition, to guarantee effective, lasting protection.

This approach allows companies to adopt high-performing, scalable and responsible cybersecurity while staying in control of innovation.

Conclusion: security, compliance and innovation

Governing the use of generative AI means integrating it intelligently into a modern cybersecurity strategy, rather than restricting it. The combination of the Cato Networks SASE platform and SASETY's support gives organisations a trusted framework to use AI while controlling its risks. This is the essential condition for responsible, secure and sustainable innovation.

☎️ A project, a question? Our experts are here to help!

Contact our experts →

💡 To discover the Cato Networks SASE platform and SASETY managed services, nothing beats a demo!