
AI-native product thinking has begun, but only if you’re ready

We’re living through a tectonic shift in how digital products are imagined, built and scaled. But here’s the hard truth: most organizations aren’t ready for what’s coming.
Immobiliare.it is the number one portal for publishing and searching property listings in Italy. We’ve spent the last few years laying the groundwork—not chasing shiny tech tools but rewiring how we think. We’re no longer just building platforms, we’re building intelligence-driven ecosystems. That evolution started with one foundational decision: to adopt Mixpanel and make digital analytics a first-class citizen across our teams.
Prior to Mixpanel, we lacked a robust digital analytics solution, finding Google Analytics too basic for our ambitions. Adopting Mixpanel was a crucial step in transforming our organization and fostering a data-informed culture, moving away from a model where data analysts were the sole owners of insights. We wanted a more democratized and accessible solution.
Let me be clear—Mixpanel wasn’t just a better dashboard. It was the enabler for everything that came next.
Data-driven culture isn’t a buzzword. It’s the operating system.
When I joined Immobiliare.it in 2021 we didn’t have a product-led culture. BI reports were disconnected, product managers were flying blind and the organization lacked a shared language around data. So we made a hard pivot: digital analytics became non-negotiable.
We embedded Mixpanel into our day-to-day, not as a reporting tool but as a thinking mechanism. Our product managers became product scientists. Conversations changed. We started speaking data fluently—at lunch, in meetings, even at parties. And that’s what it takes: full cultural saturation.
The results? We went from 20% of registered users using our app to nearly 80% in just over a year. And I can’t tell you which specific feature drove it. Because that’s not the point. Outputs are irrelevant. Outcomes are everything. Mixpanel helped us stop obsessing over vanity metrics and start understanding real behavioral shifts.
From tracking events to shaping behavior
One of our biggest wins came from connecting the dots between online and offline behavior. Like many marketplaces, a chunk of our commercial activity wasn’t even happening on our platform—it was going through the sales team. Invisible to our product teams. Useless for iteration.
But with Mixpanel’s Data Warehouse Connector, we ingested offline data into our event schema in weeks, not years. Suddenly, we could model churn risk, predict renewals, and proactively influence the behavior of real estate agents—our revenue engine. That kind of visibility used to be a dream. Now it’s table stakes.
AI isn’t a feature. It’s a new product paradigm.
Here’s where it gets truly disruptive. Our latest MVP—a dual-agent system built without engineers—uses AI to qualify buyer intent and trigger personalized outreach when agents don’t respond. It’s already delivering 90%+ email open rates and materially improving match rates. No roadmap backlog, no six-month engineering cycle. Built in two weeks and live in production.
Where we once built wireframes and RFPs, we now develop fully working applications (both frontend and backend) during the discovery phase. Engineers no longer receive static requirements; they get live, functional apps. Their job is no longer to interpret specs, but to scale working solutions.
This shift changes everything: speed, cost, risk, and predictability.
On one hand, we’ve become far more efficient in validating our product–market fit, because we test real usage, not assumptions. On the other, we’ve drastically improved the handoff to development: input structures, logic and output formats are already implemented in the MVP, making specifications much less ambiguous.
And Mixpanel remains critical. It lets us measure not just what happened, but why. It gives us confidence that the probabilistic systems we’re designing are actually moving the needle.
The future belongs to probabilistic product teams
The world is getting more ambiguous, not less. If your team still treats measurement as a hygiene step, or AI as a side quest, you’re falling behind.
The next generation of product teams will be built around three pillars:
- Universal data fluency—no gatekeepers
- Analytics-native infrastructure—from day one
- AI-powered MVPs—not as tech experiments, but as shipping products
If you’re not designing for probabilistic outcomes now, you’re already behind.
Mixpanel gave us visibility. AI gave us acceleration. But culture made it possible. And that, I believe, is the real unlock.
For more details, watch my talk from MXP London here.