A hotel butler who never sleeps, never takes a break, and can juggle four times the workload of a human counterpart sounds like the fever dream of a general manager drowning in staffing headaches. But that's exactly what Axelrod, a two-person startup out of San Francisco, is selling to luxury properties across three continents—and it claims to be doing it without touching a single line of code in their existing software systems.
The company, which emerged from Y Combinator's Summer 2026 batch and went public with its offering in August, has embedded autonomous AI agents inside Michelin Key and Marriott Luxury Collection hotels. These agents handle everything from guest check-ins and room assignments to rate adjustments and maintenance coordination, all by watching and mimicking how hotel software works rather than integrating with it directly.
It's an audacious workaround to one of hospitality's thorniest problems: the industry runs on a patchwork of legacy systems that rarely talk to each other, and persuading hotels to rip out their property management platforms is a nonstarter. Axelrod's solution is to teach AI to use the software the way a person would—clicking, typing, reading screens—using computer vision and browser automation.
"With computer- and browser-use, our agents drive the hotel's existing stack … with no … APIs, or integrations," co-founder Saman Sayahpour wrote in the company's launch post. The approach lets Axelrod sidestep the fragmented API landscape that has left countless hospitality tech ventures struggling to scale.
Under the hood, the system relies on what the company describes as two layers. A memory layer stores standard operating procedures and guest histories. An orchestration layer routes tasks to specialized agents assigned to roles like front desk, housekeeping, or revenue management. The agents start under close supervision, Axelrod says, and gradually "earn" autonomy as they prove reliable. Whether that autonomy can be trusted long-term with high-stakes guest interactions remains an open question.
Both founders bring unconventional backgrounds to the venture. Sayahpour, who dropped out of Carnegie Mellon's computer science program after being admitted to study mathematics at the University of Vienna at age 14, grew up working behind the reception desk of a Vienna hotel. Co-founder Adrian Stoica spent his childhood cleaning guest apartments in Berlin and later pivoted to music production, researching AI Awareness and co-producing records that racked up more than 40 million streams at Sony Music, according to their Y Combinator profile.
That blend of hospitality experience and technical ambition appears to have given them a sense of where the industry might be headed. Stoica attended HITEC 2024 in San Antonio in June, where much of the chatter had shifted from basic chatbots to agentic AI capable of executing complex workflows autonomously, based on his LinkedIn activity and conference recaps.
Axelrod has not named its hotel clients publicly, though it joined the California Hotel & Lodging Association and Boutique Hotels of California as a vendor member in August 2026. Its Y Combinator directory page claims deployment across three continents, while the launch post mentions four—one of several small discrepancies that suggest the company is still firming up its messaging.
The startup is far from alone in chasing this opportunity. At least three other companies from Y Combinator's Summer 2026 batch are building AI operations tools for hotels. Lance says it supports more than 50 properties across Marriott, Hilton, and Hyatt brands using similar "Computer Use Agent" technology. Conduit and Zaplar are targeting hotel operations as well, while voice-agent startup Riviera joined the accelerator's Winter 2025 batch.
Outside Y Combinator, established players are making moves. Lighthouse announced a "Revenue Agent" in February. Canary Technologies launched a "Hospitality AI Agent Studio" in March. WorldVue unveiled its Atlas operational intelligence layer in June. Ameniti, SOYL, and Alfred Hospitality also sell AI operations platforms, according to company announcements.
All of which raises the question Stoica himself posed on LinkedIn in July: "Everyone wants to deploy AI agents, but the question … is how to … do that at scale."
Axelrod's answer—autonomous software that operates invisibly inside the tools hotels already use—may prove elegant or fragile depending on how well it handles edge cases and system updates. The company has disclosed no funding beyond Y Combinator's standard investment and listed a team size of two as of September. For now, it's a lean operation betting that the future of hospitality doesn't require hotels to change their software, just who—or what—is using it.
