Lens 02 · Installation

Two weeks on site.
Then it runs itself.

No tape on the floor, no reflectors on the racking, no civil works. The installation is a survey, a map, a mission build and a safety test — and the same five steps every time.

Below is what each step is, roughly how long it takes, and who is doing the work.

2–4 weeks typical No floor markers Your floor keeps running
Step 03 · mission buildCode-Less editor, sped up
The sequence

Five steps.
Same order, every site.

Timings are typical for a single-application deployment on one floor. Fleet size, WMS integration and door or lift control are what push a project from two weeks toward four.

01

Survey the floor

Someone walks the route with a laser measure. Aisle widths, doorway clearances, ramp gradients, floor surface, charging position, network coverage and the places where people and forklifts cross. The output is a measured plan, not an impression.

Day 1–2
02

Map it with SLAM

The robot is driven or drives the route once and builds the map itself. Keep-out zones — glazing, mesh, dock edges — are marked at this stage, while somebody is standing in front of them and can point at the problem.

Day 2–3
03

Build the mission

Pickup points, drop points, routes, traffic rules and handshakes are assembled as nodes in the Code-Less editor. No code, but it is still engineering: the sequence has to survive a station being blocked, a load being wrong, and a person standing in the aisle.

Day 3–5
04

Validate the safety

Stop distances measured, fields tested at each speed, the actual load carried — not a demo box. Emergency stop, obstacle intrusion, and recovery from a blocked lane. This is the step that produces a signed acceptance document.

Day 5–8
05

Go live and train

Operators are trained on the exception paths, not the happy path. Spares are handed over, the support channel opens, and the first weeks are watched. The floor should not notice the day the robot started.

Week 2–4
Evidence · go-live

This is the screen
on day one.

The Code-Less Monitor is the live view of the floor: every robot, the step it is executing, and the step before it. If a robot stops, this is where you find out why — not from a phone call an hour later.

Code-Less Monitor · sped up per-robot step, step type, sequence, coordinates and log lines
Missions

Every flow on one canvas

Each mission is a node graph. The monitor shows them running side by side, so a stalled branch is visible as a stalled branch rather than a lower number.

Robots

Step-level, not summary-level

Current step, step type, sequence, coordinates, last completed step. Enough to tell "waiting at a station" apart from "stuck".

Logs

The reason, in words

Log lines name the action and the target — checking a stack, docking a pick. Written for the person on the floor, not only for the engineer.

Before we arrive

What we need
from you.

Required

  • A floor plan — even a PDF of the fire evacuation drawing is enough to start.
  • A named safety contact — the person who signs the acceptance test.
  • Charging location — a power point where the fleet can park, ideally near the work, not in a corner.
  • A real load sample — the actual pallet, tote or bin, at the actual weight.

Helps a lot

  • Wi-Fi coverage map — dead spots are the most common surprise on go-live day.
  • Traffic rules — one-way aisles, forklift zones, shift change times.
  • WMS or ERP contact — if the loop needs to close into an existing system.
  • Door and lift control details — the interfaces differ by building and by age.
The honest part

What actually
goes wrong.

Deployments slip for boring reasons. Every one of these has delayed a project; none of them is a robot problem, and all of them are cheaper to find during the survey than during go-live.

  • The floor plan is out of date. Racking moved two years ago; the drawing did not. Hence the laser measure on day one.
  • Wi-Fi does not reach the far aisle. The robot is fine; the mission download is not.
  • The load is different on the day. Overhang, wrap, weight distribution. Acceptance testing uses the real thing.
  • Nobody agreed who owns the exception. When a station is blocked at 3 a.m., does the robot wait, reroute, or call someone?
  • Shift change is a crowd. Throughput measured at 2 p.m. is not throughput at 6 a.m.
  • The integration was scoped last. WMS handshakes have the longest lead time of anything on this page.

A two-week install is not a two-week promise. It is what the five steps take when the survey was honest and the integrations were scoped early. When they were not, the schedule says so before the purchase order, not after.

Book the survey first.

Before a fleet size, before a quote — a walk of the route and a measured plan. Everything after that is arithmetic.