An integrator friend once joked: “Clients remember the robot price and forget everything that makes it move.” Behind the joke is the field truth—many projects do not die on insufficient payload. They die in week one: not enough IO, gripper arrives late, PLC handshake changes three times.
We roughly split integration cost into above and below the waterline. Above is equipment: arm, controller, teach pendant, baseline cabling. Below is everything that makes the arm repeat reliably on your shift.
Below the surface, layer one: interfaces and wiring
Cobots promise “plug and play,” but “play” often means the arm jogs when powered. “Plug into your line” is a different job: PLC handshakes, MES signals, e-stop chains, safety relays, tool power, vacuum manifolds—every wire has an owner.
The SME trap is assuming an integrated controller box solves logic. It saves floor space, not responsibility. Who triggers start? Who counts good parts? Who loads the next program at changeover? Those questions do not always add hardware lines, but they always add field hours.
Layer two: end effector and process proof
The arm arrives; the gripper is still on a truck—industry punchline and the number-one schedule killer. Worse: the gripper arrives, but close time is 1.4 s and your 6 s takt is dead on arrival.
Process proof is not “pick once and call it done.” It is worst case: thinnest part, oiliest surface, fastest shift change, longest tool extension. Skip any item on the list and acceptance becomes “one more tweak.”
Layer three: people and organization
Who is trained? Can the line lead teach a small offset? Who resets at night? Who approves spare parts? Is documentation current—and in the language the night shift actually reads?
Plants underestimate “who watches month one.” If it is always remote integrator support, cost does not vanish; it deferrals and returns as service fees. Internal capability is capex; outsourced capability is subscription. Both work—choose at charter time.
Layer four: downtime window and opportunity cost
Three days to mount, two weeks to tune—on paper that is labor. Off paper: a half-month of lost orders, overtime that cannot catch peak, an audit while crates are still open.
Opportunity cost rarely appears on quotes, yet owners feel it first. That is why “do it in slow season” always sells: calendars as risk management.
Making the hidden explicit
We are not trying to scare you off integration. We want one table before PO that names the below-waterline work:
Mechanical: flange, fixture mass, extension, guarding
Electrical: IO list, e-stop, PLC handshake sheet
Process proof: critical pose video, worst part, takt broken down through EOAT moves
Organization: who teaches, who first-articles, who stocks spares, support boundary after integrator leaves
Time: downtime window, rollback plan if the pilot fails
Once the underwater work is named, the debate shifts from “is the arm expensive” to “is this station worth doing now”—which is the adult version of automation planning.
If you only have an arm quote, pause. Ask for the below-waterline list. If the seller cannot answer, that is not a discount. It is risk not yet priced.
