Pre-Run Checks Before an Autonomous Scrubber Starts: A Checklist

What has to happen before an autonomous scrubber starts?

Before an autonomous scrubber leaves its dock, three things need a short look: the floor, the machine and the route. The floor has to be swept and cleared of hazards. The machine needs its squeegees and brushes inspected and its sensors cleaned. The route has to be confirmed as clear of people, carts and loose cords.

None of this takes long, and skipping it is where most avoidable trouble starts. A debris pile left behind by a sweep can jam a squeegee. A smeared sensor can leave a robot blind to a pallet jack in its path. A cord lying across an aisle can be missed entirely. Vendor guidance from Tennant and SoftBank Robotics points the same way, even though the exact steps differ by model.

Why does the floor need a full sweep first?

Tennant advises pre-sweeping the full cleaning path, not just picking up large debris. According to Tennant, dust builds up on most floors within hours, so a floor that looked clean at closing may not be clean when the scrubber starts. Debris left behind can lodge in the scrubber head or squeegee assembly, causing machine damage or streaking.

Mats come up too. Tennant advises removing mats before scrubbing, and it says to look for protruding hazards such as sockets and grates, which can damage brushes and squeegee blades. A sweep is also the moment to spot a wet patch, a spilled cart load or a gap in the route that needs a cone.

How do squeegees, brushes and sensors get checked?

Tennant's T7 daily checklist calls for checking squeegees for damage, wear and deflection adjustment, and for checking brushes for string or plastic wrapped around them. A Tennant 5680 checklist also calls for squeegee wear checks. These checks catch the faults that show up as streaks on the floor and wrapped brushes on the machine.

Sensors get the same attention. Tennant's AMR quick start guide says to visually inspect all sensors and wipe them with a microfiber cloth before starting an autonomous route. Smudges are easy to miss on a dim overnight floor, so the wipe belongs in the routine rather than in a repair ticket.

Warehouse pre-shift templates, which come from a template site rather than a manufacturer, add battery condition, tank levels, leaks and basic control function to the same walk around. They suit a shift lead writing a routine for the first time, though the manufacturer manual should still set the final list.

What can a scrubber fail to see on the floor?

SoftBank Robotics' guidance says some floor objects, including slippers, cords and retractable electrical outlets, may not be detected as obstacles, so it advises removing them before teaching a route. The same logic applies on every run after that, because a cord that was clear on Monday can be back on the floor by Tuesday night.

Tennant states that its robot cannot detect objects less than 4 inches from the floor and will run over them in autonomous mode. That figure comes from Tennant's vendor guidance, so check the wording in the manufacturer document for your model before quoting it in a site briefing.

A pre-run walk therefore looks low as well as high. Scan for cords, hoses, low chair bases and loose floor mats near the route, and lift or relocate them before the machine leaves its dock.

How should a route be confirmed before a run?

Tennant advises teaching routes when the area is clear of obstacles and people, which can mean teaching outside normal cleaning hours. Its best-practices guidance also recommends smooth, wide turns, avoiding narrow aisles and no reversing. Its area fill guide says to clear the route of debris and, where possible, temporary obstacles.

Teach-and-repeat systems start with a manual pass. Hako describes the operator driving the full intended route once before autonomous runs, in its Scrubmaster brochure. Tennant notes that a route may need re-teaching when the space has changed enough that the scrubber struggles to navigate and calls for frequent assists, which makes a confirmed route a moving target rather than a one-time task.

Not every vendor takes this approach. Avidbots describes its platform as using real-time obstacle avoidance and dynamic path planning, in contrast to teach-and-repeat, a claim that was not independently tested as of November 2026. Monster Janitorial, in a vendor blog, argues teach-and-repeat is less flexible in changing spaces and that unexpected obstacles can interrupt cleaning. Both sides are making vendor arguments, so a site should test either approach on the actual floor.

What goes on a pre-run checklist, in order?

A checklist only works when it runs in the same order every night. The sequence below puts the floor first, because debris and mats affect every step that follows.

Keep the list posted at the dock and log each run. A spot that keeps needing a second pass usually points to a route or layout problem, not a careless sweep, and the log makes that visible before the floor pays for it.

• Sweep the full cleaning path and lift mats
• Scan for sockets, grates and other protruding hazards
• Check squeegee wear and deflection, and brushes for wrapped string
• Wipe every sensor with a microfiber cloth
• Move cords, hoses and loose objects off the route
• Confirm the route is clear of people and temporary obstacles
• Check battery condition, tank levels and leaks before departure

Where does a lease fit into the overnight routine?

Service Robot Co. is an OEM-neutral integrator. We pick robots across manufacturers, then lease, deploy, integrate, train and service them through a nationwide US engineer network, so the same pre-run discipline applies whichever scrubber a site chooses. Our robot leases run on a 3-year term with one fixed payment, quoted after a site assessment.

For night shift operations, the checklist is the part that does not depend on the machine. Overnight floor care with no operator on site only holds up when the sweep, the inspection and the route check all happen before the scrubber moves. A site assessment is a practical moment to agree who completes each step, how long it takes and where the log lives.