The four-axis capability scorecard
Score every vendor on four independent axes — mechanical design, embedded electronics, ROS 2 software, and commissioning — and weight by how much your program needs each. A vendor at 9/10 on software and 4/10 on commissioning will deliver a beautiful repo and a robot that doesn't work at your facility. Most Indian robotics vendors are strong on two axes and weak on two. Pick the one whose strong axes match where your team is weakest.
Red flags in the first meeting
Five signals that move a vendor down the scorecard: (1) the pitch is on slides, not on a robot — capable vendors will show you a working system on a call. (2) they can't name the senior engineers who will work on your program. (3) their case studies are 'lab demos' rather than client deployments. (4) they quote without asking about your site conditions — floor material, lighting, network, integration targets. (5) commissioning is described as 'support' rather than a scoped deliverable. Any two of these and you keep looking.
ROS 2 maturity check — five concrete questions
Ask: (1) Which ROS 2 distribution do you target for new programs, and why? (Correct answers: Humble LTS or Jazzy LTS, with a rationale tied to release-end dates.) (2) Show me your simulation environment — Gazebo, Isaac Sim, or custom? (3) How do you handle DDS configuration for multi-robot or constrained network scenarios? (4) What does your CI / CD pipeline look like for ROS 2 packages? (5) When was the last time you contributed to or filed an issue against an upstream ROS 2 repo? Vague answers to these are diagnostic — capable teams answer specifically and with examples.
The pilot-first contract structure
Structure the engagement as a fixed-scope, fixed-price phase 1 pilot — typically 8–12 weeks, 15–25 percent of total program budget — that delivers a working demonstrator on a representative subsystem. Use the pilot to evaluate the vendor under realistic constraints (your CAD review process, your facility, your stakeholders). Only commit to the full program after the pilot lands. The 25 percent you spend on a pilot saves 50–70 percent on average across programs that would otherwise need a vendor change at the halfway point.
Contractual clauses that matter
Six clauses we recommend in every Indian robotics contract: (1) IP ownership clearly assigned to the buyer (default in India is ambiguous). (2) Source code delivery with documented build instructions at every milestone — not just at the end. (3) Commissioning hours specified as a separate line item with a defined facility location. (4) Acceptance criteria tied to objective metrics (cycle time, success rate, MTBF), not subjective demos. (5) A 60-day post-acceptance bug-fix window. (6) A change-order process with named approver on both sides. Vendors who push back on any of these are telling you something.
How Yantrix scores ourselves
Honest self-assessment because we'd ask you to do it of any vendor: mechanical 8/10 (in-house CAD, FEA, manufacturing coordination), embedded electronics 7/10 (custom PCBA through Indian houses, firmware on STM32 and Jetson), ROS 2 software 9/10 (Nav2, MoveIt, custom drivers, simulation in Gazebo and Isaac Sim, full CI/CD), commissioning 8/10 (on-site or remote, with named engineers on every program). We're strong on the software-and-software-integration axes and we're investing on the electrical side. Whether that matches your program is the right question to ask.
