Nederlands
Deutsch
We are living in what you might call the Ford Model T era of artificial intelligence: the moment a technology is embraced on a massive scale.
Not the phase of the very first experiments, but the moment a technology is embraced on a massive scale. AI use is growing explosively: people enthusiastically try out new applications, while others still stand on the sidelines sounding warnings.
The comparison with the car fits remarkably well. Horses instinctively find their way back to the stable; a car must always be actively driven. And before the car was considered safe, decades of adjustments were needed: seatbelts, traffic rules, crash barriers, driving licences.
We are at that same point now with AI. The first pioneers have already driven the new vehicle, corners are being cut and there are still hardly any seatbelts in place.
Accelerated adoption: AI for everyone
Within organisations, adoption is faster than ever. Not only innovators, but precisely ordinary employees seize the opportunities AI offers. And who can blame them:
- Work that used to take hours can now be explained in natural language to a system that understands you.
- Need an email? Name the subject and AI writes it.
- Need an image? Describe what you want and you have it within seconds.
But that convenience comes at a price.
The risk: exposure of confidential data
Every question and every answer is shared with an AI system. And that is precisely where the danger lies:
- Confidential information can unintentionally become public or part of the knowledge pool of a public AI model.
- Unprotected data can end up in the public domain this way.
- This calls for policy. This calls for guardrails.
Compliance: problems often arise unwittingly
An employee who quickly has a report rewritten via a public AI tool does not automatically think of regulations. Yet this can lead to:
- Breaches of the GDPR,
- Violation of internal guidelines,
- Fines or legal risks.
Not out of bad intent, but out of a drive for productivity.
Security: new risks and new attack vectors
On top of this, security risks arise:
- AI browser extensions that do not meet internal security requirements,
- Applications that siphon off data unnoticed,
- New forms of attack such as prompt injections.
Controlling AI use must therefore be integrated into the existing security architecture, not as a standalone tool, but as part of a whole.
The reality check: AI is already being used
Employees already use public AI tools such as ChatGPT, Claude and Gemini today. Often without formal approval or oversight. Not to circumvent rules, but to do their work better and faster. Any organisation that wants to control AI must acknowledge this reality.
Control begins with insight
Without visibility into data flows and application use, every measure remains reactive. Anyone who wants to regulate AI must first understand:
- Where data goes,
- Which tools are deployed,
- What risks that creates.
Transparency is the foundation. Control begins with measuring.
AI in the contact center: innovating without data leaks
In contact centers, the pressure to innovate is high. AI offers enormous opportunities:
- Real-time assistance,
- Intelligent routing,
- Automation that increases both efficiency and customer satisfaction.
But speed must never come at the expense of control. Shadow AI arises here when employees deploy tools themselves for transcriptions or summaries. This creates uncontrolled data flows of highly sensitive customer information. Safe AI use means that AI functionality must be embedded in enterprise solutions.
Solutions such as those from Five9 show that innovation and compliance can indeed go together. By implementing AI in an integrated and managed way, security stays assured and scalable innovation arises without governance risks.
Momentum EMEA
