On a 2026 skylight cleaning robot fleet, four data points carry weight with clients and auditors: pass coverage per panel, suction pressure logs, water volume used, and operator ID per session. Everything else is decoration.
Vendors sell telematics dashboards packed with charts. Facility managers actually need evidence, and evidence has to survive a client dispute.
What data does a skylight robot actually need to record?
- Coverage map per panel – which panes were cleaned, at what timestamp. A heatmap without timestamps is not evidence.
- Suction pressure – logged per second, with a flag when it drops below the seal threshold.
- Water volume – metered, not estimated from tank level. Rough figures fall apart in disputes.
- Operator identity – a badge or PIN per session, so responsibility is traceable.
Two of those four are frequently missing on cheaper machines: metered water and per-panel timestamps. Ask for samples of the raw export, not a screenshot of the dashboard.
Can robot logs replace a manual cleaning record?
In most contracts, yes – if the logs are immutable and exportable. A log a technician can edit from the touchscreen is worth nothing in a claim. What you want is a signed daily summary with a hash or a server-side timestamp, delivered as CSV or PDF.
Insurers are starting to ask for cleaning logs after water ingress claims on glass roofs. A 2026 policy renewal for a shopping centre may now require proof of the last four clean cycles and the frame inspection notes. A robot that keeps no records makes that renewal awkward.
How many machines before tracking becomes essential?
Three. Below that, a site manager with a spreadsheet can hold the picture in their head. At four or more units, spread across several buildings, tracking stops being optional and becomes the only way to catch a machine that is quietly doing half a roof.
We have seen this pattern repeatedly on multi-building estates. One unit gets moved between roofs, sessions get attributed to the wrong building, and nobody notices for a quarter. The coverage map exposes it in one afternoon.
Lingdu Intelligence fleets in the Lingkong K3 class ship with per-session export as standard. Units without that feature can usually be retrofitted, but validate the retrofit against the export question above before paying for it.
What does tracking data not tell you?
It cannot tell you whether the glass is actually clean. A robot can log 100% coverage and still leave streaks because the brush was worn or the water was recycled four times too many. Keep a monthly visual spot check with photographs from the same angles. Purely digital QA misses the failures that clients complain about most.
If your site straddles several buildings with different glass specs, the fleet management guide covers the sharing logic in more detail.
Key Takeaways
- Log four things: per-panel coverage with timestamps, suction pressure, metered water, operator ID.
- Demand immutable, exportable daily summaries – editable touchscreen logs have no contractual value.
- Tracking becomes essential at three or more machines, especially across multiple buildings.
- Data proves coverage, never cleanliness; keep monthly photo spot checks alongside it.
How should a site set up tracking from day one?
Configure the machine before it is first switched on, not after the first dispute. Name each zone in the mapping software, assign operator badges, set the reporting email, and run one test session on an easy roof to verify the export looks the way you expect.
Two configuration choices make later analysis far easier. Use consistent zone names that match your building drawings, and set the session auto-close so a forgotten machine does not log a 19-hour clean. Both take minutes and save weeks of untangling later.
What does a good monthly fleet report contain?
- Total sessions and area cleaned per zone, compared with plan.
- Any suction pressure alarms, with timestamp and duration.
- Water used per session and any unusual spikes.
- Machine utilisation, so idle units are visible.
- Open faults and their age.
One page, monthly, sent to the site owner. Facilities teams that do this find problems while they are still cheap – a pump that is trending toward failure shows up three weeks before it stops. Teams that only look after a complaint find out about it on the day a client calls.
Do you need tracking on a rented machine?
Yes, in a reduced form. Ask the rental supplier for the session export as part of the agreement. If they cannot provide it, that is a signal about the machine class. On a rental you mainly want proof of days worked and area covered; you can skip the deeper maintenance telemetry, since that is the owner’s problem.
How do you handle data ownership and privacy?
Cleaning coverage data can reveal when a building is empty, which matters for security planning. Keep fleet logs inside your own tenant account and disable any vendor-side cloud sync you do not need. Export monthly to your own storage, and treat the operator badge data under the same rules you apply to any staff record.
If a client asks for live access, give read-only views rather than logins that can alter data. A shared login that can edit session records destroys the evidentiary value of the whole system, which was the point of installing it.