
Matteo Risso

Connect in LinkedIn
Matteo is Lead Product Manager at WeRoad, where he leads cross-functional squads working on lower-funnel conversion, ancillaries, and post-booking experiences across European and US markets, and led the product side of WeRoad's US launch. Over 8+ years in product he has built digital experiences for travel, e-grocery, and proptech scale-ups, after a first entrepreneurial run as founder of Makeapporter, a beauty services marketplace.
A talk with

Cassiano Vailati
Head of Technology
WeRoad
< About >
Everyone is experimenting with AI. Very few companies have turned it into the system they actually run on and even fewer are willing to show the messy middle.
Over the last two years WeRoad went from the first proof of concepts to an AI-native company, and this talk retraces that journey chapter by chapter, in the order it actually happened: the first AI products shipped to customers, AI-assisted engineering landing on every developer's desk, the company-wide rollout that put powerful tools in everyone's hands and the sprawl that came with it. That chaos is what led to the current chapter: a Company Brain that centralizes organizational knowledge, a data warehouse that gives AI real context, and a growing layer of automated processes and pipelines built on top.
For every chapter, Cassiano and Matteo share what was built and what was hard: learning curves, models that change under your feet, governance and trust at company scale. Matteo closes the loop from the product side, showing how this foundation transformed the way product managers work from prompting to shorten tasks to recurring agents embedded in the daily flow. The session ends with where the journey goes next and Q&A.
< Key learnings >
Start by shipping, not by strategizing - WeRoad's journey began with real AI products (support chatbot, search, translations, auto-documentation) and the unglamorous lessons that came with them: testing, verification, and new releases break things.
AI-assisted engineering compounds fast - What happened when agentic coding tools reached every engineer and PMs started opening pull requests: release cadence more than doubled in two years, with review overhead and cognitive overload as the real costs to manage.
Company-wide rollout creates sprawl before it creates value - When everyone can suddenly do everything, you get duplicated work, unverified answers, and a data team drowning in "but my numbers say otherwise". Governance isn't optional.
The Company Brain & Governance tools are the answer to the chaos - How WeRoad is moving from personal AI setups to a centralized brain: RAG over internal documentation and code, plus a data warehouse semantic layer that gives every AI interaction real organizational context.
A new operating model for product teams - How this foundation changed PM work end to end: from one-off prompts to recurring agents that update tasks after groomings, write issues, prototype on the design system, and open PRs.