Service Robots
Most “pilots” are a robot in the lobby for a month and a folder of guest smiles. A real pilot has one job, a named owner, and a rule for sending the machine back.
Use this for Relay on amenities, Temi on a travelling screen, or Pepper on a short greeting script — one of those, not all three. Pair it with lobby robot versus a tablet or a person and lease versus buy so you are not piloting a purchase you already made.
ⓘ Quick Answer
Run one indoor route for 30 days with a staff owner on every shift. Keep it only if the route still runs in week four without a manager hovering. Send it back if it needs rescuing more than it saves walking.
Set the pilot before it arrives
- One start point, one end point, one payload.
- A paper tally: trips completed, trips rescued, minutes saved, guest or staff complaints.
- A lease or return in writing.
- Wi-Fi, dock space, and lift access tested with a trolley first.
What to measure in weeks two to four
Week one is novelty. Weeks two to four are the product. Count rescues, charge gaps, and whether the named owner still wants the job. If only the general manager is excited, the pilot has failed even if the video looks good.
The kill rule
Write it on day one: “If we rescue it more than X times a week in week four, it goes back.” Also kill it if content is stale, if it blocks a fire route, or if staff hide it. Paying for a second month to “give it a chance” is how cupboards fill up.
💡 What to Check
Photograph the dock location and the narrowest door on day zero. Those two pictures prevent most failed pilots.
30-day pilot checklist
30-day pilot checklist
- One job, written in one sentence.
- Named owner on each shift, not “the team”.
- Tally sheet for completed trips and rescues.
- Return or lease exit already signed.
- Kill rule dated for the last week, not “we’ll see”.
Final Thoughts
A pilot is an experiment with an exit. If you cannot fail, you are not piloting — you are decorating.
Pick one service robot on Robot.help, run the route, then decide.
Frequently Asked Questions
Is 30 days long enough?
Yes for a single indoor route. Not enough for a building-wide rollout. Do not expand until week four is boring.
Should guests see it on day one?
Better to run staff-only for a few days so the first public trip is not a rescue.
What if the vendor runs the pilot for us?
Useful for setup. Your staff must still own week four, or you have tested the vendor, not your building.