What Kinetic Eye's own pages state, and the specifics an evaluation still has to ask for, written for EHS, operations and risk leads in warehousing and industrial sites. Every constraint below carries the documentation behind it, the operational impact, and the workaround.
Every constraint below comes from Kinetic Eye's own home, product, company, blog and privacy pages at kineticeye.io, and from the press release its own footer links, checked on August 14, 2026. That date is older than this page on purpose: kineticeye.io could not be reached over a verified connection on August 17, so this page carries the date the evidence was actually gathered rather than claiming a check that did not happen. Most of what follows is an absence on a public page rather than a finding about the product, and this page says so each time rather than reading a gap as a no. Spot AI sells a competing platform, so nothing here rests on an anonymous source or an aggregated user rating.
The team, the intent and the business case are published. The mechanics of the product are not, and the site's own footer adds a question about the brand itself.
A constraint list is only useful next to an honest account of the product.
Six people appear with their roles, from Josh Butler as chief executive and Michael Richmond, PhD as chief technology officer through customer success, computer vision engineering, product engineering and a robotics and vision advisor. Kinetic Eye states its team have built AI products for Tesla, Facebook, Google and Microsoft and developed technology to keep people safe in self-driving cars. A buyer weighing a small vendor is really weighing the team, and this page answers that directly.
Kinetic Eye publishes that by reducing your risk, you can improve your mod rates and reduce payroll and training expenses, which puts the argument on the experience modification rate that sets a workers' compensation premium rather than on a detection count. Most platforms in this category sell alerts and leave the finance conversation to the buyer. This one opens with it.
Chris Ashford, managing partner at Same Day Delivery, and Dan Weiss, chief operating officer at Unchained Logistics, both speak on the home page, and an SAP logo sits under a Partnered with SAP heading. The company page states plainly that your data is just that, your data. The blog is real work too: six posts on back injuries, slips and falls, ergonomics, OSHA violation types and safety signage, citing Bureau of Labor Statistics figures rather than the vendor's own.
Each one is a consequence of how much this vendor publishes rather than a defect. What matters is whether it collides with how your evaluation has to run.
A Brand Update link in the footer of kineticeye.io opens a press release dated April 18, 2022 stating that Kinetic Eye is becoming CompScience, that its identity, logo and domain name have been updated, and that the company was entering the insurance industry with AI powered workers' compensation insurance. It also states the company was founded in 2019. The kineticeye.io pages describing the safety platform are still live and unchanged in that respect.
This is a fact about the company published by the company, and it is the first thing to settle rather than a mark against the product. A quote, a support agreement and a data processing agreement all have to name an entity, and a procurement file that cites pages belonging to a former brand will be sent back at the legal review.
Ask in writing which legal entity would be the counterparty, what the product is called today and where its current documentation lives, then evaluate that documentation rather than these pages. Confirm the support arrangement and the roadmap owner in the same message, and attach the answer to the procurement file.
Automated risk identification, powered by computer vision is the published capability, and nothing is listed behind it. The published outputs are a web hazard dashboard, alerts to your inbox, trends over time, as many users as you need and an implementation Kinetic Eye works through with you. So what reaches a person is documented even though what triggers it is not.
You cannot check personal protective equipment, forklift proximity, falls, restricted zones, spills or ergonomic risk against the pages before a call, which means the first meeting is spent establishing scope rather than testing fit. Every other vendor on a typical shortlist publishes at least a list, so the comparison sheet has one column that cannot be filled in.
Send your hazard register and ask for a written yes or no per line, plus a recorded clip of each detection running on a real site. Treat the scope as an open question for the vendor rather than as a documented gap, and put the written answer beside the published lists from the rest of the shortlist so the comparison is like for like.
No protocol, model list, analog route or on-site hardware requirement appears on the home, product, company or privacy pages, and no statement of where video is processed or stored appears either. The documentation covers what the software shows a user, not what it needs from the estate to show it.
For an operations team with a decade of mixed hardware across three warehouses, this is the single question that decides feasibility and whether the project carries a hardware line at all. A program approved on a software budget can turn into a camera replacement program once the answer arrives.
Send the camera list with make, model, firmware, resolution and frame rate, and ask for a written compatibility answer per line before any pilot is scoped. Ask two follow-ups in the same message: whether anything runs on site, and what network path the vendor needs from each location.
Your data is just that, your data is stated once on the company page. The privacy policy behind it is dated February 17, 2020 and governs website visitors rather than the video platform. Nothing states how long footage is kept, where it lives, who at the vendor can reach it or what happens to it at the end of a contract.
An ownership sentence is a good starting position and it is not an answer a regulated site can give its own auditor. A logistics operator with an insurer requirement, a union agreement or a customer audit clause needs a period, a location and a deletion commitment in writing.
Request a data-handling pack by name: the retention period, the storage location and region, the sub-processor list, the support-access policy, the end-of-contract deletion terms and any certification report. Ask for a data processing agreement covering the platform rather than the website, since the published policy covers the latter.
Deployment is published as easy to implement, with Kinetic Eye stating it will work with you to execute quickly. No timeline appears in days, weeks or any other unit, and no statement of what an implementation includes appears beside it. That is what the page says rather than an inference about how the work goes.
On a multi-site warehouse operation the implementation is usually the largest line after the software itself: site surveys, network changes, mounting work and the internal time to review what the dashboard produces. None of that is visible until a scoping call, so an early business case carries a blank where its biggest variable belongs.
Ask for a written statement of work before committing: what the vendor does, what your team does, how many days per site, what any on-site work costs and who owns the network changes. Then hold the first site open as a paid pilot with a defined exit, so the scope is proven on one location before it is repeated across twelve.
The same five constraints in one view, sized to paste into an evaluation document.
Swipe the table sideways to see every column.
Kinetic Eye data comes from Kinetic Eye's own home, product, company, blog and privacy pages at kineticeye.io and the press release its own footer links, checked on August 14, 2026, the last date the site could be read. Gaps are marked as not publicly specified.
If the person signing is accountable for a mod rate, and the argument that lands internally is that fewer incidents cost less to insure, Kinetic Eye frames the problem the way that buyer already frames it, on the home page rather than three calls in. Most of the constraints above are about published detail rather than capability, so a buyer who is comfortable establishing scope in conversation, and who has the time to do it, is not blocked by any of them.
Spot AI answers the feasibility questions before the first call rather than during it, which is the practical difference here. Any ONVIF or RTSP IP camera works at full functionality and legacy analog comes in through the Intelligent Video Recorder, so a warehouse estate assembled over a decade can be assessed from a camera list rather than a site visit. Full-resolution video stays on the IVR in the building and only event metadata crosses the network, which is a documented architecture rather than an ownership sentence.
The detection list is published too. 15+ pre-trained Video AI Agents cover vehicle break-in, fire, cash register theft, after-hours intrusion, personal protective equipment, forklift near-miss, falls and crowding in hazard zones, and Iris builds anything else in natural conversation in about eight minutes. The platform is SOC 2 Type II, NDAA-compliant and HIPAA-aligned, and the response runs through talk down, strobes and horns on standard speakers already on site rather than stopping at a dashboard.
Decide first whether the business case is a safety metric or an insurance metric, because that decides who sponsors it. Then run whatever survives against the cameras already on the wall.
None of that makes Spot AI the right answer for every buyer. Kinetic Eye ties its outcome to the mod rate that sets a workers' compensation premium, and Spot AI publishes no insurance premium or experience modification claim at all: its evidence is safety, investigation and operations outcomes reported by named customers. Kinetic Eye also names its team on a page, which a buyer weighing a small vendor may value more than a feature list. Spot AI is a video AI platform rather than a risk-financing product, and access control and sensors arrive through integrations. A constraint list earns its keep by holding each platform's shape against your starting point, and here that starting point is usually who in the building is paying for the outcome.
A live pilot on your cameras answers in a week what a spec sheet cannot.
Customer-reported outcomes from named Spot AI customers.
Unique Industries covers more than a million square feet with a three-person safety team, catching near misses and falls on the cameras already installed.
Silver Bay Seafoods replaced fragmented legacy camera systems across 22 locations, including remote Alaska facilities, and lifted operational efficiency 15%.
Blackmon Oil runs a vehicle loitering filter to keep parking lots clean and safe at its around-the-clock stores, with a smaller overnight crew.
"Having cameras that talk back to you changes everything. Spot AI has become a real part of our safety team."
Five, and four of them are about what the pages do not publish rather than about what the product does. The footer links a press release announcing a change of name and business; no hazard class is named behind the published capability; camera compatibility and architecture are not publicly specified; the data-ownership line carries no retention or storage terms; and the implementation is described as easy rather than scoped in days or deliverables. Each one is a question for the vendor rather than a documented no.
The site says it itself. A Brand Update link in the footer of kineticeye.io opens a press release dated April 18, 2022 announcing that Kinetic Eye is becoming CompScience, that its identity, logo and domain name have been updated, and that the company was entering the insurance industry with AI powered workers' compensation insurance. The kineticeye.io pages describing the safety platform are still live, so confirm which entity and which product a quote covers before anything else.
Not publicly specified. No camera protocol, model list, analog route or on-site hardware requirement appears on the home, product or company pages, and Kinetic Eye does not state either way whether an existing estate can be reused. Treat that as an absence on the public pages rather than as a documented no, and put it first on the list of questions for the vendor, because on a mixed warehouse estate the answer decides the whole project.
Automated risk identification, powered by computer vision is the published capability, and no hazard class sits behind it. Nothing on these pages names personal protective equipment, forklift proximity, falls, restricted zones or any other detection, and no accuracy, latency or coverage figure appears. The published outputs are a web hazard dashboard, alerts to your inbox and trend reporting over time, so what reaches a person is documented even though what triggers it is not.
It depends on how much has to be settled before a purchase order. If the evaluation needs the camera answer, the detection list and the architecture in writing, a camera-agnostic platform such as Spot AI documents all three: any ONVIF or RTSP camera connects as it is, legacy analog comes in through the Intelligent Video Recorder, and 15+ pre-trained Video AI Agents cover safety alongside security and operations. If the business case is the insurance line specifically, that is the one place Kinetic Eye is more direct than most of this category.