SPOKE ABOUT #3 — Vibe Coding, Agentic Enterprise & GEO: with Mathieu Morgensztern (AdaptivAI)
Podcast

SPOKE ABOUT #3 — Vibe Coding, Agentic Enterprise & GEO: with Mathieu Morgensztern (AdaptivAI)

April 15th, 2026 · 5 min read

Back to news

For this third episode of SPOKE ABOUT, Jérôme Duchamps welcomes Mathieu Morgensztern, founder of AdaptivAI — an entrepreneur with a rare background, who spent years leading major communication groups before diving into the AI tech world. A dense, precise and hype-free episode to understand why 2026 is a pivotal year.

The Spark: An "AI for Business" Master's at Oxford in 2021

Before ChatGPT, before the mass-market wave, Mathieu Morgensztern was already completing an "AI for Business" master's at Oxford. That head start allowed him to understand what many discovered too late: generative AI is not just another trend. It's a deep paradigm shift that is reshaping the very way we work.

When ChatGPT arrived, he wasn't surprised. He was ready. And he founded AdaptivAI to connect AI to the real knowledge of enterprises.

Vibe Coding: Extraordinary, but Limited Without Expertise

Vibe Coding — the ability to generate code through natural language instructions — is presented as revolutionary. And it is. But Mathieu offers a precise nuance: a poor developer using Cursor does not become a software architect.

What holds for code holds for every profession: tomorrow, we'll have vibe marketing, vibe finance, vibe HR. AI doesn't replace expertise — it amplifies it. And those who lack it have nothing to amplify.

2026: The Year of the Agentic Enterprise

This may be the episode's strongest conviction: 2026 will be the year of the agentic enterprise. Every key function — marketing, finance, HR, sales — will have its own AI agents, connected to its data, its tools and its objectives.

These agents won't just produce faster. They'll produce better: more analysis, more insights, higher quality at every step. The promise isn't speed — it's elevation.

Stop Looking for "Use Cases"

One of the most common mistakes in AI adoption? Spending time searching for use cases in the abstract. Mathieu Morgensztern proposes a radically different approach: bring your actual work from the week and test AI on it, immediately.

Adoption can't be decreed. It's lived, in daily work, on real tasks. That's how you move from curiosity to action.

Creation Remains Deeply Human

AI can execute a brief. It can generate, adapt, optimize. But finding an idea, designing a campaign, having a strong creative intuition — that remains human. Mathieu states this without ambiguity, coming as he does from the world of communication.

The real question isn't "will AI create in our place?" but "how does AI free up our capacity to create more?"

GEO, Data Sovereignty and Choosing the Right LLM

The episode also covers topics companies can no longer ignore: GEO (Generative Engine Optimization) — this new SEO designed for AI engines — the question of data sovereignty, and the criteria for choosing your LLM (carbon cost, transparency, business fit).

These are strategic decisions that, made too late, will have a real cost on organizational competitiveness.

The Quote That Sums It All Up

"AI is not just another tool. It's a new way of working."

A simple phrase, but one that captures the essential: this isn't a question of tools, it's a question of posture. Companies that have understood this are moving forward. The others are still waiting.

Listen to the Episode

Find this third episode of SPOKE ABOUT in audio and video format:

And you — is your enterprise ready for the agentic era? Let's talk about it.

Let's discuss your project