Your floor. Their robot.

Tested in simulation. Before you sign.

Merci. Verified trust for physical AI.

Book a floor-risk review
Scroll

01 — The Trust Gap

The contract is multi-year. The demo was twenty minutes.

Deploying a warehouse robot still comes down to faith. The demo, the pilot video, the uptime statistic — every piece of evidence in front of you was produced by the company that wants your signature.

Finding 01

AT-RISK

Vendor demos run in vendor conditions

Wide aisles, clean concrete, perfect light. Glare, debris, and shift-change traffic — the conditions most likely to break a robot — are the ones a demo floor never has to show you.

Finding 02

FAIL

Everyone grades their own homework

Uptime figures, pilot results, acceptance tests — authored by the seller of the machine they describe. A robot scorecard the robot-seller pays for is worth exactly as much as a home inspection the seller pays for.

Finding 03

AT-RISK

$13.8B went into robotics in 2025, up from $7.8B in 2024

Crunchbase's funding count. The fleets are arriving faster than the diligence around them — and the buyer's side of the table is still largely without an examiner of its own.

02 — The Examination

It passed the vendor's test. Now ours.

Merci is an independent examiner for warehouse robot fleets. We grade a machine none of us built, against the aisles, lighting, and dock traffic it will actually face. The vendor demonstrates; we examine. Those are different jobs — and they cannot be done by the same party.

Preparing digital twin…
Scenario 014 — pick-aisle transit
4 conditions flagged

Their demo floor doesn't have your dock-door glare.

Or your shrink-wrap debris, your aisle 14, your forklifts at shift change. What you're watching is a closed-loop examination: the robot's autonomy drives, a calibrated facility model pushes back — and the conditions a polished demo never meets are exactly the ones we run first.

03 — Method

One floor. Thousands of scenarios. One report.

An examination, not a demo. Built on NVIDIA Isaac Sim and Cosmos, run closed-loop — the robot's autonomy drives, and a model of your facility pushes back.

Scan your floor's conditions

We capture what makes your facility yours — aisle widths, dock-door light, floor surfaces, traffic patterns, the debris that actually accumulates — and build a calibrated simulation model from it.

Run thousands of closed-loop scenarios

The robot's autonomy stack drives the simulation while your conditions push back. Glare at 4pm, a pallet half in the aisle, pickers at shift change — swept systematically, not sampled by chance.

Receive ranked failures and a test plan

One report: the conditions most likely to make this robot fail on your floor, ranked by likelihood and explained — plus the physical tests to run first, in order, before you sign.

04 — Who relies on the report

One report, read from four sides of the deal.

01

Operators & 3PLs

Turn a faith-based purchase into an evidence-based one — ranked floor risks and a physical test plan, before the capex and the SLAs are committed.

02

Systems Integrators

A vendor-neutral acceptance scorecard for fixed-price commissioning — de-risk your own exposure, and walk into operator deals with independent evidence in hand.

03

OEM Autonomy Leads

Commissioned by your buyer, never by you — but the harness runs in your VPC, your stack stays in your environment, and it can surface edge cases your field data hasn't hit yet.

04

Robotics Insurers

Failure modes, recovery behavior, and environment-specific risk, documented by a party with no stake in the sale — in the language underwriting already speaks.

You'd never close on a building without an inspection.

Thirty minutes. We walk your floor's conditions, flag what's most likely to break a robot there, and scope what a full examination would cover. No obligation — and no vendor is told you called.

Book a floor-risk review

We grade robots none of us built. Paid by the buyer, never the seller — no robot-vendor partnerships, no logos to protect, no stake in whether you sign.

Built on NVIDIA Isaac Sim and Cosmos, evaluated closed-loop: the robot's autonomy drives, your simulated floor pushes back. The deliverable is risk-surfacing and a prioritized physical test plan — not a certification, and not a guarantee.