So your warehouse or production line is finally getting robots, and the integrator proposals are rolling in — one priced per cell, one priced per project, one priced per “automation journey” (whatever that is). If you can’t line up the robot integration quotes you’ve received, you can’t choose between them, and whatever you sign becomes a leap of faith with a six-figure deposit. Good news: making quotes comparable is a repeatable procedure, and you can run it without a procurement department. Here’s the whole thing, start to finish.

First, write the job down like you’d brief a new operator

Before contacting anyone, produce a two-page job description for the robots. What gets moved, lifted, welded, or inspected; cycle times and throughput targets; the layout constraints and the humans working nearby; what your existing systems (WMS, MES, PLCs) must talk to; and the one metric that defines success — picks per hour, defect rate, whatever you’d put on the wall. If AGVs are in the picture, our full guide to automating warehouses with AGVs will help you name the decisions (navigation method, fleet sizing, traffic rules) that belong in this document rather than in a sales call.

Next, decide how you’ll judge the offers — before you see any

Pick your axes now, while you’re still objective: lifetime cost (not sticker price — include integration hours, spares, service contracts), the integrator’s evidence with your task type and payloads, their service footprint relative to your facility (a robot down for a week while a technician flies in erases a year of savings), and how much of their claim set they’ll commit to in writing. Fixing the judging criteria before the first proposal arrives is the single best defence against being steered by the shiniest demo video.

Then send the same document to four or five integrators

Identical text, identical deadline, and a required answer format: costs broken out by hardware, integration labour, software licensing, and first-year service; assumptions stated; a named project engineer; references from comparable deployments. Tell every integrator that others have the same brief — it sharpens pencils wonderfully. If anyone replies with a generic brochure instead of your format, you’ve learned something about how they’ll handle your punch list later.

Borrow the comparison stage from an unlikely place

The cleanest model for what happens next isn’t industrial at all — it’s medical. On MyHospitalNow, which gathers hospital treatment bids against a patient’s one written request, the submission is free, a review pass checks the requirement, and then the patient does exactly what you’re about to do: reads several written estimates together, weighing the charge, each provider’s profile, the geography, and the trust signals (the platform is the one vouching that its hospitals are verified), and commits to nothing until one offer is formally accepted — only then are the two parties connected. Notice the order of operations: comparison is a protected stage, fully separated from both the selling and the signing. Give your robot project the same protection — a block of calendar time where the only activity is reading the bids against each other, with no integrator in the room.

Normalise the numbers before you believe them

Quotes rarely arrive on the same basis even with a fixed format. One integrator amortises software across three years, another bills it upfront; one assumes your electricians run the cabling, another includes it. Pull every quote onto one spreadsheet basis: total cost of ownership over five years, at your duty cycle, with your labour assumptions. Expect the cheapest sticker to move two or three positions once you do this — that movement is the entire reason the step exists.

Interrogate the top two, especially on the sensing stack

Shortlist two and go deep where robotics projects actually fail: perception and integration. Ask each finalist to walk you through what sensors they’re specifying, why, and what happens when dust, glare, or a mispositioned pallet confuses them. You’ll evaluate the answers better after skimming our explainer on how sensing hardware gets built into industrial robots — fifteen minutes of background turns “trust us” into a real conversation. Push on commissioning time too; optimistic schedules are the most common fiction in this industry’s proposals.

Decide once, in writing, against your own document

Hold the decision meeting with three artefacts on the table: your original job description, the normalised spreadsheet, and the finalists’ written commitments. Pick the offer that best satisfies the metric you defined on day one — not the one with the best demo, and not automatically the cheapest after normalisation. Record the reasons in a paragraph. Then, and only then, start contract and scheduling conversations with the winner, and send the runners-up a courteous note. Keep their quotes filed; they’re your benchmark when change orders appear, and the integrator market is small enough that today’s runner-up quotes next year’s expansion.

The checklist version, for your office wall

Written job description first. Judging axes fixed before proposals exist. One identical brief to four or five integrators with a required answer format. A protected comparison block — read together, normalise to five-year cost, shortlist two. Deep technical interrogation of the finalists. One documented decision against your own success metric. Every step is boring, which is exactly the point: robots are the exciting part, and the quoting process is where you earn the right to enjoy them. Run it this way once and you’ll never let a sales calendar schedule your automation budget again.