Information for journalists.
What we are, what our numbers mean, how to cite us, and what we will and will not say. Everything on this page may be quoted or reproduced with attribution.
Media inquiriesGoes to our contact form with the media topic already selected. We aim to reply within two business days.
What Robovations is
Robovations is an independent publication that classifies consumer robots by how much they actually do without a person. Every robot in the database is placed on the same five-level Autonomy Ladder™, scored against a published rubric, and reclassified when the evidence changes.
The short version, for quoting or paraphrasing: Robovations classifies consumer robots by autonomy level using a published methodology, and tracks how those classifications change over time.
The distinction that matters most in coverage is between classification and recommendation. A Level IV robot is not a better purchase than a Level III one. The level describes capability, not fit, value, or quality. We grade those separately and we never rank.
By the numbers
These figures are generated from the live database each time this page loads, so they are current as of your visit rather than as of the last time someone updated a press kit.
- I5 1%
- II159 34%
- III275 58%
- IV33 7%
- V0 0%
Level V is defined as a robot that learns new tasks in new environments without guidance. No shipping consumer product has demonstrated it, and the tier stays empty until one does. That is a finding, not an oversight.
Attribution
Attribute to Robovations. We do not offer individual bylines, named spokespeople, or personal quotes, and that applies to press coverage as much as to the site itself.
This is a deliberate editorial position rather than a media-relations preference. Every classification is the output of a published method applied to documented evidence, and the framework should outlive any one classifier. A Level III is a Level III regardless of who applied the rubric. Naming an individual would suggest the judgment rests on their authority when it rests on the rubric and the sources, both of which are public and checkable.
What that gives you instead of a quote: the full methodology, the source list behind any individual classification, and a dated reassessment trail on every robot page. If a classification looks wrong, the evidence it rests on is on the page and you can argue with it directly.
What we do and don’t do
Worth being precise about before a story runs, because the two halves are easy to blur.
- Classify every robot against a single published rubric, applied identically across the database.
- Aggregate and date-stamp public evidence: manufacturer documentation, independent teardowns, press reviews, owner reports, firmware history, and longitudinal signal.
- Reassess and revise classifications when the evidence changes, keeping the previous record visible.
- Publish the sources behind a classification and the depth tier that describes how much corroboration it has.
- Operate or test the robots ourselves. We aggregate public evidence and say so on every methodology page.
- Rank, recommend, name winners, or publish star ratings. Classification is not a purchase verdict.
- Accept products for review, payment for placement, or manufacturer hospitality.
- Certify anything. We classify. No manufacturer can apply for, purchase, or appeal a level.
The frameworks
All four are published in full, including their limitations. If you are checking whether the method holds up, these are the pages to read.
-
What can the robot do on its own?Five levels, applied identically to every robot. Benchmarked against SAE J3016, ISO 8373, ISO 13482, and ANSI/HFES 400, with the departures documented.
-
How is a classification sourced?Six evidence channels and four depth tiers. Verified requires five or six of the six channels reviewed.
-
How well does it do what it claims?Nine inputs across three pillars (capability, dependability, burden), reduced to one number out of 100 with the components always shown alongside. Cost of ownership and data handling are published beside the score, never inside it.
-
Can a buyer trust it today?A separate read on availability, support maturity, and post-purchase risk, independent of capability.
Assets and contact
If you are citing one specific robot, use the embed studio at the bottom of that robot’s page. It generates copy-paste markup for the classification mark in seal or standard form, light or dark, alongside a formatted citation. The mark links back to the assessment and reflects the current classification, so it stays right if we reassess after your piece runs.
Each also exists as a dark-background version (-dark) and a single-colour version for one-colour printing (-mono-black / -mono-white), at /assets/logo/.
For generic use, where no single robot is the subject, the five Autonomy Ladder™ level marks are below, in the seal size. Each also exists in the narrower standard size at /badge/autonomy-L{1-5}-standard-{light|dark}-embed, and both come in a dark-background version. All are SVG.