The Story of Xenia
·4 min read·Lodovico Benvenuti
Share:

The Story of Xenia

The Sublet That Started Everything

In August 2021, I faced a problem: paying rent on an apartment in Milan while studying at Cambridge. Listing it on Airbnb revealed a profitable opportunity—rental income exceeded housing costs.

A friend named Marco asked a pivotal question: "Why do we need to be there at all? Why can't we just put a lockbox on the door?" This sparked the realization that technology could eliminate manual work and enable scaling.

The Notebook Era

Early 2022 brought the first client property, managed via physical notebook. The operation quickly evolved to Google Sheets with custom scripts for scheduling and occupancy tracking. When discovering Property Management Systems, they seemed revolutionary but ultimately insufficient.

The Decision to Build

By early 2024, managing 30 apartments alongside day jobs became unsustainable. I shifted perspective from people management to engineering problem-solving, building task management systems, scoring algorithms, dispatch optimization, and chatbot automation.

Recognizing limitations, I reached out to Host AI (later Conduit), a YC W24 startup. Initial contact was rejected, but five months later, after consulting competitors like Boom and Dharma, I joined as an employee in San Francisco.

The Host AI Chapter

Building reliable AI for guest inquiries proved harder than expected—five months versus anticipated one month for knowledge base development. The experience provided foundational work for true operational coordination.

When Conduit shifted focus in late 2025, I returned to Milan with renewed purpose.

The Xenia Model

Today, Xenia manages 150 apartments in Milan, generating €3 million revenue in 2025 with only four employees: one commercial person and three rotating operational staff.

The philosophy: "if a computer can do it better, a human should never touch it." Every process follows the progression: Tracking → Standardization → Optimization → Automation.

The vision mirrors Waymo's autonomous driving approach—automating hospitality operations' coordination and execution.

The Stack

PMS: Kross Booking - Italian market-specific system handling reservations, invoicing, and compliance.

Guest Messaging - AI handles 90%+ of inquiries without human intervention.

Field Operations - Breezeway plus custom backend generating intelligent daily planning.

Pricing - PriceLabs dynamic pricing generates 15-25% revenue increases.

Access - Hermes custom system using Shelly devices for remote unlocking.

Accounting - Custom scripts for Italian invoicing and reconciliation.

What Automation Actually Means

A broken air conditioner at 2pm triggers instant classification (HIGH urgency, SIMPLE complexity). The system identifies Alessandro's location 2km away, recalculates his afternoon route, provides briefing with access codes and part information, and auto-generates guest updates. Human approval requires one decision point.

Evening algorithms optimize next day's field operations using urgency scoring, damage coefficients, inspection recency, route clustering, and guest constraints.

Checkout triggers automatic cleaning assignment to nearest available cleaner with access codes and deadlines. Quality issues become maintenance tasks slotting into following dispatch automatically.

The Real Unlock

Automation means freeing humans for novel problem-solving, relationship building, and judgment calls—not eliminating people. The shift transforms Property Manager 1.0 (reactive firefighting) to Property Manager 2.0 (experience engineering and execution supervision).

What's Next

By end of 2026, Xenia plans operating on a "single AI-governed backend"—a digital twin representing every conversation, asset, task, issue, and workflow. A "gold rush" surrounds building first AI agents handling property management end-to-end: coordination, orchestration, and execution.

The Advice

  1. Track first - measure before improving
  2. Standardize second - establish repeatable processes
  3. Optimize third - eliminate unnecessary steps
  4. Automate last - only when processes are clean and predictable

"Never copy-paste a check-in instruction again."


Related reading: The Future of Hospitality explores where the industry goes next, and The Great 2026 Consolidation examines why fragmented tech stacks are dying.

Lodovico Benvenuti
Lodovico Benvenuti

Co-Founder at Trellis. Scaled STR operations to $2M+ revenue and built the AI backbone behind Conduit before founding Trellis.

© 2026 Trellis Tech, Inc. All rights reserved.