The Setup
YYForce Inc. (NASDAQ: YFOR) has acquired 20 wearable data collection systems — pairing head-mounted egocentric cameras with synchronized wrist-mounted cameras — and is deploying them inside Novotel Singapore on Stevens to capture first-person recordings of experienced housekeeping staff at work. The footage will be annotated and processed to support humanoid robot learning research, including work with vision-language-action (VLA) models. It marks one of the more concrete steps any commercial robotics firm has taken toward automating back-of-house hospitality operations using real-world, task-specific training data.
What Gets Captured
The data collection program targets three workflow clusters: bed and linen preparation (bed-making sequences, linen handling, pillow arrangement); towel and amenity organization (folding, placement, and guest-room supply replenishment); and selected cleaning activities (surface wiping, tool and materials handling). The goal is to document hand-object interactions and natural task variation in a live operating environment — the kind of contextual nuance that synthetic or lab-recorded datasets rarely replicate. YYForce is clear that collecting demonstrations does not, on its own, establish autonomous capability or commercial readiness; the recordings must still be reviewed for quality, annotated, and adapted to specific robotic hardware before any task evaluation begins.
"Access to real operating environments and experienced service teams is an important part of our robotics development strategy," said Mike Fu, Chief Executive Officer of YYForce. "This initiative will help us document how housekeeping tasks are performed in practice and assess how those demonstrations can support humanoid robot learning."
Why It Matters for Operators
For operators tracking foodservice automation and robotics trends, the Novotel pilot is a useful bellwether. Humanoid robots have long been demonstrated in controlled factory settings, but hospitality and facility-service environments — irregular room layouts, variable guest behaviors, mixed surfaces — present a far harder problem. By grounding model training in actual hotel workflows rather than simulated environments, YYForce is betting that demonstration-based learning can close that gap faster than traditional programming approaches.
The broader energy-and-sustainability pressures hitting hotel operators are also relevant context: labor costs for housekeeping have climbed sharply in markets like Singapore, where staffing constraints in service industries remain acute post-pandemic. If humanoid platforms can eventually handle repetitive linen and surface tasks reliably, even at partial scale, the operational math changes for large-property managers.
Data collection will follow agreed hotel procedures and applicable privacy requirements, including participant notification, guest privacy protections, access controls, and permitted use limitations — all to be confirmed before recording begins, the company says. YYForce combines workforce solutions and integrated facility management with its robotics development arm, giving it direct access to the operating environments needed to validate the technology outside a laboratory.
Written by Michael Politz, Author of Guide to Restaurant Success: The Proven Process for Starting Any Restaurant Business From Scratch to Success (ISBN: 978-1-119-66896-1), Founder of Food & Beverage Magazine, the leading online magazine and resource in the industry. Designer of the Bluetooth logo and recognized in Entrepreneur Magazine's "Top 40 Under 40" for founding American Wholesale Floral, Politz is also the Co-founder of the Proof Awards and the CPG Awards and a partner in numerous consumer brands across the food and beverage sector.
