Post-market surveillance has the same shape as patient access: most inputs are routine, a few are serious, and the system is judged entirely by whether it catches those few.
Post-market surveillance (PMS) under MDR Articles 83 to 86 is the manufacturer’s system for collecting and analysing data on a device once it is on the market: complaints, service reports, literature, registry data and user feedback. Its outputs are the PMS report for class I devices and the periodic safety update report (PSUR) for higher classes, and its findings feed back into the risk file and clinical evaluation.
Inside that stream sit the serious incidents that Article 87 requires to be reported to the competent authority within fixed deadlines, the shortest being two days for a serious public health threat. So the first thing the system does with any complaint is screen it for a possible serious incident, with the threshold set deliberately low, and send any hit to a person with the complaint attached. Coding and trending come after.
The vigilance decision is always a person’s. A model cannot decide that an event is not reportable, and in what we build it is never asked to.
email, call, field, distributor
runs first, threshold set low
IMDRF codes suggested
in your complaint system
PMS report, PSUR draft
possible serious incident
Under MDR Article 85, manufacturers of class I devices prepare a post-market surveillance report summarising PMS data and any preventive and corrective actions, updated when necessary. Class IIa, IIb and III devices instead need a periodic safety update report under Article 86, updated at least annually for the higher classes. Both draw on the same PMS system, which is where AI drafting helps.
It can do the reading, structuring and coding, and it can screen every complaint for a possible serious incident before anything else happens. A complaint handler confirms each record, and the decision on whether an event is reportable is always made by a person. The system is tuned to over-escalate, because a missed incident is the failure that matters.
Article 87 sets maximum periods after the manufacturer becomes aware: two days for a serious public health threat, ten days for a death or unanticipated serious deterioration in health, and fifteen days for other serious incidents. Those clocks are why we put incident screening first and route possible cases to a person immediately rather than batching them.
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.
We start by testing incident screening on complaints your team has already assessed, including the reportable ones.
30-minute technical call · no deck