The wrong robot is often selected long before the demo. The team starts with a product category, compares headline specifications and only then asks whether the machine fits the actual workflow.
A better process reverses that sequence. Define the operating problem, measure the site and set the evidence required for a decision. Product comparison comes after that.
01 / Task
Define the job in operating terms.
“We need a cleaning robot” is a category, not a requirement. A useful brief describes the work that should move from people to the robot and the result the business expects.
Replace a broad request with a measurable one.
That sentence already gives the shortlist a shape. It identifies the floor type, operating window, coverage requirement and acceptable intervention rate.
02 / Baseline
Measure today before promising tomorrow.
A robot can only improve a process that has a baseline. Record the current labour time, completed area or tasks, interruptions, quality checks and recurring exceptions.
Headline throughput is still useful, but it should be treated as one input. Route complexity, refill time, charging and recovery from exceptions determine the real output.
03 / Site
Make the site part of the specification.
Two identical robots can perform very differently in two buildings. Check the environment before narrowing the product list.
- 01Surfaces and thresholds
Map floor materials, slopes, transitions, door tracks and gaps.
- 02Traffic and operating windows
Record people, carts, peak periods and shared routes.
- 03Infrastructure
Confirm charging, water, drainage, connectivity and lift access where required.
- 04Local ownership
Name the person who resets, inspects and escalates when the normal workflow breaks.
04 / Pilot
A demo shows the robot. A pilot tests the operation.
A useful pilot runs in the real environment, against a baseline, for long enough to expose repeatable problems. The plan should state the pass criteria before the robot arrives.
Baseline, routes and success criteria
Setup, training and exception log
Coverage, intervention and quality
Result against the agreed baseline
05 / Commercial model
Choose who owns the equipment and who owns the operation.
The product is only part of the decision. Purchase, rental and RaaS place different responsibilities on the customer and the robotics partner.
| Model | Best when | Customer manages | Partner manages |
|---|---|---|---|
| Purchase | Long-term use and internal capability are clear | Ownership and daily operation | Deployment and agreed support |
| Rental | Timing or demand is temporary | Daily use within the rental scope | Equipment and agreed service |
| RaaS | The customer wants an operating outcome with less internal overhead | Site access and local coordination | Deployment, orchestration, maintenance and agreed operations |
06 / Shortlist
Finish with three credible options, not twelve brochures.
Each shortlisted robot should clear the task, site and commercial requirements. The final comparison can then focus on meaningful differences: expected output, intervention load, deployment requirements and total operating responsibility.
- 01
One safest fit
The model with the fewest unresolved operating risks. - 02
One stronger-performance option
Higher potential return with clearly stated assumptions. - 03
One alternative operating model
A different balance of equipment, service and internal workload.