Inspection records and photos hold early warnings that nobody has time to read across. We build systems that find hazards and patterns in them and put each one in front of the person responsible. The system never declares anything safe.
All Mining & Resources →
A mine generates inspection records every shift: workplace examinations, ground control checks, conveyor and guarding inspections, mobile equipment pre-starts. In the US, MSHA requires workplace examinations by a competent person; in the EU, Directive 92/104/EEC sets minimum health and safety requirements for surface and underground mineral extraction. The records exist. The pattern across them, a ground condition noted four times in a month in the same heading, is what gets lost.
Safety is also where a wrong answer is not symmetrical. A system that flags too much costs some review time; a system that misses a hazard and implies an area was checked is dangerous. So the design rules are fixed: the model only raises findings, it never clears an area, never closes a finding and never replaces the examination. A missed hazard in testing is not a metric to improve; it is a stop.
form, notes and photos from the shift
hazards in images and text
against history by area and equipment
with evidence, location and priority
confirms, acts, closes with evidence
imminent danger indicators
Uses include computer vision on cameras and inspection photos to spot hazards such as missing guards or people near moving equipment, analysis of inspection records to find recurring problems, proximity detection, and predictive maintenance to catch equipment failures early. In every case it supports the people responsible for safety; it does not replace workplace examinations or their judgement.
No. Workplace examinations are a legal duty carried out by a competent person, and the examiner's judgement on site is the control. AI can make the examiner's records more useful, highlight hazards in photos and show patterns across many inspections, but it cannot declare a working area safe.
We do not build worker scoring or emotion recognition, and the EU AI Act prohibits emotion recognition in the workplace. Vision models look for conditions such as missing guarding, spillage or missing PPE in inspection photos. Any camera-based use is designed with your works council or employee representatives and your data protection lead, under the GDPR.
What we built, what stage it is at, and a link to the live thing wherever there is one.
yourOS is an operating system with an AI at its heart. You talk to it like a person, out loud or typed. It makes your calls, plans your days and works on your projects with a crew of agents, on your own private server. Our own product, built from the first commit; the waitlist is open.
Website, backend and app assets for Sonora, a free AI sound-therapy app on iOS and Android published by Cognitive AI.
Brand, website, prototype, apps on five platforms, the AI backend and the algorithm-intelligence pipeline for Podswap, a social growth engine that runs without a human in the loop.
In closed beta: Roxy, a private AI assistant with permanent memory running on infrastructure the user owns. Brand, site, product, backend and conversational design.
Live at coherence.com: brand identity generated with AI agents we directed, the motion system and website for the group behind Sonora, AboutAI, WakeUp and MudShadows. Agent work remains in build.
Live at domainbooth.com for DomainBooth, a premium domain brokerage: a new animated mark, every page redesigned and rebuilt with designed mobile layouts, and a custom CMS for its sales record, featured names, press and enquiries.
Start with six months of inspection records. We will show what recurs, and where.
30-minute technical call · no deck