Hi, we’re Lodovico Benvenuti and Jan Sahagun, and we’re building Trellis, AI agents that run operations for short-term rental property managers.

Our Launch Video: https://www.youtube.com/watch?v=Zki0_icE3Xw

The problem

Short-term rental operations break as companies scale.

As you add listings, coordination explodes across guest messaging, cleaning, maintenance, and homeowners. Work ends up spread across 5 to 10 disconnected tools, duck-taped together by burnt-out teams, constantly firefighting issues, with information inevitably falling through the cracks.

Quality drops, reviews follow, and growth stalls.

This is because managing a property is complex. Let’s say the guest calls about a broken shower. The operator has to identify the reservation in the PMS, search a separate knowledge base for context, collect photos, and create a task in another system.

Then check a workforce tool for availability, reprioritize tasks, dispatch a technician, follow up with the guest, and later log costs, update accounting, and notify the homeowner. All this across 5-10 systems of record.

The core issue is simple: no software has full operation context to control operations end to end.

What we built

Trellis provides AI agents that run operations across the entire stack.

The agent communicates with guests, dispatches and coordinates cleaners and maintenance, files claims and costs, and follows up until issues are fully resolved.

It connects to PMSs, messaging channels, and internal tools, but more importantly, it maintains full operational context: who the guest is, what property they’re in, what the issue is, what’s already been tried, who is available, what tasks are in progress, and what needs to happen next.

Instead of passing information between tools, the agent owns the entire workflow from first message to resolution and reporting.

When agents take over operations, execution becomes consistent and verifiable. That shifts control of the guest experience from marketplaces back to operators, who can now deliver reliable service at scale, build real brands, and own demand.

Our background

Before Trellis, Lodo built AI customer service agents for short-term rentals at Conduit (YC W24). He also scaled a property management company to ~$2M revenue and 150+ listings. With trellis we are now running their operations with 1 person.

Jan spent years in the trenches of the STR industry: leading early sales at multiple AI tools for property managers, building an STR tech media platform (STR Tech Report), consulting across the space, attending 20+ industry events, and speaking with thousands of property managers along the way. That journey earned him deep trust and relationships throughout the STR world.

If you run a property management company, we’d love to show you Trellis.