Generative AI Is not a wave; it’s a decision shift

Generative AI is often described as the “next technological wave.”
It is not.

For years, my work focused on exposing data to support decisions: building streaming dashboards for operator eSIM products and then adding layers of intelligence such as inventory forecasting and fraud detection. The goal was clear: transform data into actionable insight and create measurable business value.

Today, I build systems that don’t just inform decisions – they help make them.

This shift forces a fundamental rethink. It’s not a faster version of the old job; it changes how knowledge is produced, how decisions are shaped, and how systems interpret intent.

This is not simply automation. It is the emergence of systems that actively participate in creating knowledge and shaping decisions – and therefore, in creating value.

From data to decisions: a new value chain

For years, “data-driven” meant structuring information, building models, and supporting human decisions. Value came from better visibility and improved recommendations.

Generative AI extends this by making knowledge directly usable. It reduces the gap between data and action. Across enterprise use cases, this transformation is already visible:

• Knowledge becomes probabilistic rather than deterministic
• Expertise is distributed between humans and models
• Trust must be continuously constructed, not assumed

The result is a shorter path from data to value. Systems don’t just surface insights; they translate them into outputs that shape decisions.

In sensitive domains such as payments, identity, and cybersecurity – this is not just an efficiency gain. It redefines how decisions are made, controlled and audited under strict regulatory and security constraints.

The question is no longer how to extract value from data but how much decision making we entrust to systems.

The illusion of intelligence

GenAI can simulate reasoning but carries no accountability. It produces coherence, not certainty.

Outputs are often trusted because they “look right,” even when they are not grounded in correct data or the full context. Real-world examples already show the consequences: incorrect chatbot guidance leading to financial compensation, or fabricated legal references accepted because they appeared credible.

The issue is not that these systems fail; it’s that they fail convincingly. Creating value with AI requires disciplined interrogation:

• What assumptions shape the model?
• What is optimized and what is ignored?
• Where is human validation non-negotiable?
• When does “likely” replace “verified”?

From generating outputs to designing trust

Futurism is no longer about predicting what technology will do. It’s about designing the conditions under which new systems remain trustworthy in real-world environments.

From building and deploying these systems, one reality quickly becomes clear: the model itself is only a small part of the challenge. Most of the complexity lies in everything around it – data quality, integration, security, latency, and scalability.

Designing trust in Generative AI requires:

• Translating technological potential into operational reality
• Challenging overconfidence in automation
• Contributing to governance, not just innovation
• Connecting disciplines across technology, regulation, and human factors

Responsibility by design

If AI systems are part of the decision chain, responsibility cannot be an afterthought. It must be designed into the system from the start.

This includes transparency, traceability, robustness, and human oversight; especially in critical flows. It also requires humility: acceleration is not always progress. A code review that moves from two human checks to a single AI-assisted check may appear more efficient, but it can quietly remove safeguards, despite giving the illusion of completeness.

Value is not created by speed alone; it is created by reliable outcomes.

What it means to be a futurist in the age of Generative AI

Being a futurist today is ensuring that acceleration does not outpace comprehension.

It requires three disciplines:

• Learn – continuously update understanding
• Interrogate — challenge systems and outputs
• Connect — bridge domains to reveal blind spots

The futurist is no longer observing change, but shaping how it becomes usable, safe, and meaningful.

The future will not be defined by who uses GenAI first, but by who integrates it responsibly. Because ultimately, the real question is not what AI can do with data, but how to turn those capabilities into decisions that we can defend.

By Pierre Hanne, GenAI Domain Lead, IDEMIA Secure Transactions