Conversations, Coffee Machines, and AI: Lessons from Paris and Lausanne
Have You ChatGPT’d Yourself Yet?
You’ve probably Googled yourself at some point — but have you ever ChatGPT’d yourself?
Try it. Just type your name into a GenAI tool and ask:
“Who is this person?” 
Someone asked me this last week during a short road trip between Paris and Lausanne, where I attended two great conferences: VivaTech and the Microsoft AI Tour. The “have you ChatGPT’d yourself?” question was just one of many fun conversations that came up. And, as usual, I took notes — lots of them.
With a little help from GenAI (ironically), I’ve shaped those notes into something semi-decent and hopefully useful 🙂 So if you want to hear my thoughts on GEO as the new SEO, the dawn of real AI adoption, the future of AI UX, and why pricing GenAI is a CFO’s worst nightmare, read on.
SEO Is Turning into GEO
Search engine optimization has been one of the most important digital growth strategies of the past two decades. Companies that cracked it early gained a massive edge. Google evolved from “organizing the world’s information” to becoming the most profitable ad machine in history — and the rest of us adapted.
Now we’re seeing version 2.0.
Tools like ChatGPT, Perplexity, and Copilot are changing how people find and interact with information — and with that, how companies are discovered. And just like SEO birthed a playbook, Generative Engine Optimization (GEO) will require one too.
Smaller companies are already jumping in. They’re more agile, quicker to experiment, and they recognize that this is still open territory — and to some extent, a zero-sum game. If they claim a niche now, they might hold it for good.
What’s interesting is that with GenAI, a five-person marketing team can now deliver personalized, scalable customer experiences that used to require 50 people. That’s game-changing — and they know it.
Unlike SEO, which was about winning Google’s algorithm, GEO is about earning trust inside conversational agents. New game, new rules.
GEO is still in its infancy, but the space is moving fast. Emerging tools like Clay and Copilot are already helping companies search for leads and automate outreach and initial customer interactions. As mentioned, the playbook is still being written — and those bold enough to experiment while it’s still forming stand to reap serious rewards.
For more, I recommend:
GEO Is the Next SEO (Forbes Agency Council)
From Proof-of-Concepts to Proof-of-Value
Since ChatGPT launched, we’ve seen a flood of MVPs, demos, and “look what we built” slides. Many didn’t make it past the PoC stage.
But something has changed.
Two and a half years later, I’m finally seeing real traction. Enterprises are building AI Centers of Excellence, aligning strategies, and delivering. Startups have gone through their early pivots and are zoning in on where real value lies.
One trend was everywhere at VivaTech: almost every startup pitched a “Copilot for XYZ.” From surgeons to electricians (yes, really), the agentic model — AI that helps people make better decisions and automates the boring stuff — is spreading fast.
In B2B, teams are looking at every role and asking:
“Can we build a Copilot for this?”
On the enterprise side, companies are reviewing their vendor landscape and asking which suppliers are integrating agents into their solutions. Startups, meanwhile, are scanning for gaps in that same landscape — and trying to become the preferred partner.
There’s a race on. And the winners will be those who move fast and go beyond just wrapping ChatGPT in a sleek UI.
Compared to just a year ago, the ecosystem is much more ready — investors, customers, and enterprises are far more fluent in GenAI. The adoption curve is bending fast.
Coffee Machine Insights: What Comes After the Screen?
Every conference has its sweet spot. For me, it’s the coffee machine — my go-to camping spot for spontaneous chats and accidental insights.
That’s where I met a VR engineer demoing a crime-solving experience. A few minutes into the conversation, I found myself rethinking how we’ll interact with AI in the near future.
I’ve always assumed that AI UX would evolve differently in B2B and B2C:
- B2B: Enterprise tools from players like Microsoft, SAP, and Salesforce will become the main gateways to GenAI.
- B2C: Voice interfaces will dominate, especially with younger generations always wired in with headphones.
But then came the twist.
He explained how recent chip advancements are enabling GenAI to run directly on smart glasses — with low latency and no need for constant cloud or phone access. In short: blinking could replace scrolling.
That’s both exciting and slightly weird.
Imagine trying to scroll your favorite social app with a blink, while the person across from you thinks you’re winking at them. Are you flirting or doomscrolling? Who knows.
Previously, costs were high and the hardware was too bulky for daily use. But now we’re seeing AI-ready chips that are over 25% smaller and 50% more power-efficient — making truly wearable GenAI not just possible, but inevitable. If this trend continues — and Ray-Bans can actually do the trick — we might be looking at the first GenAI interface people actually want to wear.
And here’s the kicker: some predict these hurdles could be resolved within just a few years. If that happens, this won’t just be a new UI — it could be the biggest user experience shift since the iPhone in 2008 and ChatGPT in 2022.
Learn more:
Qualcomm Targets Smart Glasses Market (Investor’s Business Daily)
Pricing AI: Still an Art, Not a Science
We’ve seen three major shifts in how B2B software is priced:
- Per-user licensing – One license per person
- Usage-based models – Pay for compute time or API calls
- Token-based pricing – Welcome to the GenAI era
In this new model, cost depends on how much text, data, images, or video your AI system processes — measured in tokens.
A chatbot summarizing legal documents will use far more tokens than one answering yes/no questions. In B2C, the challenge is even trickier. A retail bank offering a GenAI assistant must estimate not just how many users engage — but how deeply they’ll engage.
A “What’s my balance?” query is cheap. A request for a 100-year investment trend analysis… not so much.
For vendors and customers alike, this is a CFO’s headache. You can’t just set a flat fee when the cost per interaction varies so wildly. You need to test, iterate, and learn — and eventually move toward predictability.
We’ll get there. But for now, pricing GenAI is still a mix of art, science, and guesswork.
Dive deeper:
Executive Guide to AI Agent Pricing (Forbes Business Council)
Wrap-Up
Let me know what you think — and if you do ChatGPT yourself, share what it says.
Hopefully something flattering. Or at least mildly entertaining. 
