Ask any skylight cleaning robot supplier five things before you sign: suction performance on wet glass, spare part lead times, model-specific CE documentation, throughput with stops included, and who repairs the machine locally. Vague answers are a warning.
Buyers rarely get burned by the robot itself. They get burned by what surrounds it, spares, support and honest numbers. These five questions surface the real risk before money changes hands.
1. How does the robot hold suction on wet glass?
This is the single most important question, and the easiest to dodge. Ask for the suction figure measured on wet glass with a dirty skirt, not on a dry clean lab panel. Rain residue, condensation and dust film are what your roof actually presents.
If the answer is a peak dry-lab number with no wet figure, you are being sold a spec, not a solution. Magnetic and vacuum systems both exist, but for glass it is suction, and suction depends on the skirt seal being perfect.
2. What is the real lead time on wear parts?
Get it in writing. Suction skirts, brushes and filters are the parts you will order repeatedly. If they ship from overseas with a 3 to 6 week lead time and you hold no spares, every failure costs a month of lost cleaning.
Ask specifically: which parts should I stock, and what is the delivery time to my country? A supplier who cannot name the parts is not set up to support you.
3. Can you show model-specific CE documentation?
Not a generic certificate, not one for a sister model. The CE declaration of conformity must name the exact model. Ask for it with the quote, alongside the IP rating evidence and battery transport paperwork.
| Document | Why it matters |
|---|---|
| CE declaration (by model) | Site and insurer compliance |
| IP rating evidence | Proves wet-work suitability |
| Battery transport docs | Avoids customs delays |
| Local-language manual | Operator safety and training |
4. What throughput do you guarantee with stops included?
Quoted m2/h usually excludes refills, battery swaps and repositioning. Ask for the figure that includes them, and ask what battery and tank assumptions sit behind it. Then compare that to your own roof dimensions and working window.
A supplier confident in the machine will give you a conservative real-world number. One who only repeats the marketing figure is telling you something about the after-sales support too.
5. Who repairs it when it fails locally?
Machines fail. What matters is the recovery path. Is there a local service partner, a remote diagnostic option, or do you ship the robot back to the factory? A 4 week round trip for a sensor fault is a real operational cost.
Ask for names, not promises. A distributor you can call beats a portal you cannot reach.
Bonus question: is anything on this roof beyond the robot?
Good suppliers will tell you where their machine stops. A roof with pitches over 35 degrees, glass under 200 m2, or tight heritage frames may not suit a robot, or may need only a partial deployment. A supplier who says the robot fits everything has not looked at your building.
What answers should make you walk away?
Any supplier who claims the robot fits every roof, quotes peak suction with no wet figure, has no named spares path, or cannot produce a model-specific CE document is telling you how the after-sales relationship will go. The machine may still be decent, but the support around it will not be.
How do you compare two shortlisted machines fairly?
Put both against your own roof, not a brochure. List your glass area, pitch, dirt load, access route and working window, then ask each supplier for numbers against that list. The one who engages with your building honestly is usually the one who supports it honestly.
What should be settled in the contract?
- Spares list and delivery times
- Training days and who delivers them
- Model-specific CE and IP documentation
- Service response time and fault recovery route
- Warranty terms on battery, motors and pumps
How do you run a fair trial before committing?
Ask for a short on-site trial on your own glass, in your own weather, with your own operator. Six to eight hours on a real roof tells you more than any specification sheet. Watch how the robot handles your worst panels, not your best ones.
If a supplier will not trial, that is information too. A vendor confident in the machine and its support will usually welcome the chance to prove it on your building.
What does a good supplier answer look like?
Short, specific, and slightly unglamorous. Names of parts, lead times in weeks, a real wet-glass suction figure, and a clear statement of what the machine cannot do. A supplier who volunteers a limitation has usually thought about your building, and that is the person you want supporting a machine you will run for years.
Finally, revisit the questions after delivery, not just before. The answers that mattered at purchase, lead times, documentation, support, become the things you live with every month. A supplier who answered well and then delivered on those answers is the one worth keeping for the next roof.
Key Takeaways
- Demand suction figures measured on wet, dirty glass, not lab peaks.
- Get spare part lead times in writing and stock the main wear items.
- CE declarations must name your exact model, not a generic range.
- Ask for throughput with refills, swaps and repositioning included.
- Confirm a real local repair route before buying, not after failing.
Run these five questions past us and see how we answer. Get in touch here.