It's 2 a.m. at a Marriott somewhere in America, and the front desk agent is fielding a call about late checkout. After hanging up, they click through a property management system that looks like it was designed before the iPhone existed—because it probably was. Nested menus, mouse-driven workflows, the works.
Now imagine that scene playing out across thousands of hotels, night after night, with staffing levels that never quite recovered from the pandemic. That's the operational headache a Y Combinator-backed startup called Lance thinks it can solve, though not in the way you might expect.
As of February 2026, Lance claimed it had signed up more than 50 hotels spanning Marriott, Hilton, and Hyatt properties. The pitch? AI agents that don't require sleek new integrations or modern APIs. Instead, they watch the screen, move the cursor, and navigate the creaky legacy systems already in place—essentially mimicking what a human employee would do, pixel by pixel.
The company calls this approach "vision-based computer-use agents," and it's banking that the concept can unlock automation in an industry where decades-old software is not an exception but the rule. Whether that gambit works at scale is another matter entirely.
Most hotel AI platforms need integrations to function—clean APIs, structured data feeds, the kind of technical infrastructure that simply doesn't exist in many properties. Lance's founders are sidestepping that entirely. Their agents observe interfaces the way a person would, clicking through ancient menus and filling out forms in systems that might predate social media. It's a technical workaround born of necessity, and one that could define the company's trajectory.
The startup launched three products in February 2026, each targeting a different operational pain point. The Receptionist AI Agent handles calls, texts, emails, and messaging apps around the clock, fielding guest requests and escalating when things get complicated. The Sales AI Agent tackles inbound inquiries and RFPs, pulling availability and pricing from hotel data without human intervention. Then there's Lance Pad, a tablet-first operations hub designed to replace walkie-talkies and coordinate housekeeping, maintenance, and service teams in real time.
According to Lance's website, setup takes less than an hour, and the system integrates with existing hotel technology stacks. The company hasn't named specific PMS or CRM vendors publicly, though. Customer testimonials on the site claim call-volume reductions exceeding 30 percent, though these remain unverified by independent sources. That's not unusual for a young startup still finding its footing, but it does make the traction claims harder to verify.
The founding team brings a blend of AI infrastructure chops and hospitality operations experience. CEO Caleb Chan previously led product and engineering at Silicon Valley Commerce and spent time at Google and Salesforce. COO Gavin Brennen worked on AI agents at Domo and held positions at SmartSuite and Zeus. CTO Gatik Trivedi has a more varied résumé—he founded and exited a medtech venture called MedTec Innovate, conducted speech and language research at Stanford, built agent infrastructure at GI Partners, and worked on large language model evaluation at Knowtex, another YC company from the Summer 2022 batch.
The team remains small. LinkedIn lists the company at two to ten employees as of this month, and Lance is actively recruiting founding engineers. It's the kind of scrappy early stage where everyone wears multiple hats and the product roadmap is probably still being debated over lunch.
In its Y Combinator launch post around February 7, 2026, Lance said it was "working with some of the leading hotel groups in the US operating 50+ hotels" across Marriott, Hilton, and Hyatt properties. The company's website displays logos for those brands, along with AC Hotels, Courtyard, DoubleTree, Fairfield, and Hilton Garden Inn. But no hotel names, no case studies, no client references have been disclosed publicly.
Brand logos alone don't confirm commercial relationships—plenty of startups have stretched that convention—and Lance hasn't published the kind of detailed metrics or customer stories that would settle the question. It's worth noting, perhaps more than the founders might prefer.
The timing is both opportune and challenging. Lance is entering a market that saw a flurry of activity throughout late 2025 and early 2026. Wyndham Hotels & Resorts deployed an agentic guest service platform with 250 AI agents in production last November, citing reductions in handle time. Langham Hospitality Group launched three AI agents for multilingual guest support in December. Canary Technologies announced a global AI Voice rollout with Wyndham the same month. Duve, another player in the space, raised a $60 million Series B in December to scale its unified AI-driven guest experience platform.
PolyAI has been deploying enterprise voice assistants in hospitality for a while now, with a Golden Nugget case study published just last month. And Lance shares its Y Combinator batch with Riviera, another startup building multilingual AI voice agents for hotels. That's a lot of competition for a small team with a sub-hour setup promise.
Broader infrastructure plays are also in motion. Sabre Hospitality launched updates to its SynXis Concierge.AI in mid-2025 and announced an end-to-end agentic AI travel platform with PayPal and Mindtrip in February. Lighthouse introduced Connect AI around the same time, offering a data connectivity layer for AI agents across property management systems. The Hotel Yearbook 2026, published this month, discusses the rise of agentic AI in operations—a sign that the concept is moving from vendor pitch to industry-wide conversation.
Still, the operational pressures that Lance is targeting are real. Hotel operators have been dealing with persistent labor shortages for years now. A February 2025 survey by the American Hotel & Lodging Association found that 65 percent of hotels reported staffing shortages, a figure that has been widely cited throughout 2025 and into this year. Those constraints create an opening for automation that can handle routine tasks without requiring new hires or additional training overhead—assuming the technology actually works as advertised.
Lance is positioning its agents as "your hotel's best employees," a framing that emphasizes augmentation over replacement but still leans into the operational gap created by staffing pressures. It's a careful tightrope walk, common in enterprise AI pitches, though the actual labor implications remain murky.
The company hasn't disclosed funding amounts beyond its Y Combinator backing, and the Winter 2026 batch Demo Day timeline has passed. Lance is still in the early stages—no named customers, no published case studies, and metrics that rely on customer testimonials rather than third-party validation. There's a lot left to prove.
The technical bet on computer-use agents to navigate legacy systems is genuinely interesting, especially in an industry where decades-old software is common and modern integrations are sparse. Whether that approach scales across the wildly diverse hotel tech stacks in use today, and whether the claimed 50-plus hotel deployments translate into long-term contracts and expanding footprints, remains an open question.
For now, Lance is racing to demonstrate that AI agents can do more than answer phones—they can operate the creaky, mouse-driven infrastructure that still runs much of the hotel industry. It's an audacious premise. Whether it's a viable business is another matter, and one the market will decide soon enough.
