Run the Table-Turn Test
Pudu Robotics, Bear Robotics, and Keenon Robotics build indoor platforms for food running, dish return, and front-of-house presence. All three work in a quiet dining room, which tells an operator very little.
The Bot Scout table-turn test is the commercial question underneath: does the deployment increase completed covers during the busiest ninety minutes. Outside peak the floor is not the bottleneck, so a robot that performs beautifully at 3pm is decorative.
The hard constraint is aisle geometry. Chairs move, guests stand, and servers cut across the room, and a robot that must wait for a clear path loses precisely when speed matters most.
| Job | Where it helps | Where it fails | Measure |
|---|---|---|---|
| Food running | Long runs from kitchen to far tables | Tight aisles at peak | Server walking time removed |
| Dish return | Heavy bussing loads | Loading takes longer than carrying | Bus tubs cleared per hour |
| Host and guidance | Waits and seating flow | Guests who ignore it | Time to seat at peak |
| Drinks and refills | Repeat trips to one station | Spill risk on turns | Refill trips removed |
| Marketing novelty | Short-term attention | Wears off quickly | Not a durable business case |
Count Handoffs, Not Trips
Every run has two ends. Somebody loads the robot in the kitchen and somebody unloads it at the table, and if those two actions together take longer than carrying the plates, the robot has added work — the handoff-count discipline in the service robots hub.
The food-specific timing constraints, including plate temperature and the last-50-feet handoff, are covered in the food delivery robots guide.
Pilot during a real service with a full room. A vendor demonstration in an empty restaurant measures navigation, not usefulness.
- Measure completed covers at peak, before and during the pilot.
- Time loading and unloading at both ends of a run.
- Walk the aisle at peak and measure the narrowest point.
- Plan charging location and cleaning between services.
- Decide who rescues a blocked robot mid-service.
Staff, Guests, and Upkeep
Servers decide whether the robot is used. If it is unreliable at peak, staff route around it within a week, and the asset becomes furniture. Train the people who will actually work alongside it, not just the manager who signed the contract.
Guests vary more than vendors admit. Some enjoy it, some ignore it, and some will not unload their own plates, which puts the handoff back on a server.
Cleaning, charging, and daily checks are new tasks in a business that has no spare labour at peak. Assign them explicitly or they will not happen.
Bottom Line
Restaurant robots are worth it when they increase covers at peak and shorten handoffs. Measure table turns during real service, and assign the charging and cleaning work explicitly.
Measure completed covers during your busiest ninety minutes, before and during any pilot.
FAQs
Do restaurant robots actually save labour?
Only if the loading and unloading at both ends take less time than carrying the plates. Otherwise the robot moves effort rather than removing it.
What is the biggest obstacle for a restaurant robot?
Aisle geometry at peak. Chairs move, guests stand, and servers cut across the room, so a robot that waits for a clear path loses time exactly when it matters.
Are robot hosts worth it?
They can help with waits and seating flow, but the measure is time to seat at peak. Novelty value fades quickly and is not a durable business case.
How should a restaurant pilot a robot?
During a real service with a full room, measuring completed covers at peak, handoff time at both ends, and how often staff must rescue a blocked unit.