Controlled use of AI in your quality system
AI tools can support quality work when they are under control. I help you write the procedure for using AI, validate the tools you use and train your team, so the AI supports your work without creating new risks.
Three situations where this works best
People use AI tools for drafting, reviews and analysis, without a procedure or validation.
You see where AI could save time, and you want to introduce it under control.
Software used in the quality system has to be validated for its intended use, and AI tools are included.
Your team already uses AI
You want to start carefully
An auditor will ask
You don't know which AI tools your team uses, or for what.
No procedure says what AI may and may not be used for.
AI tools touch quality records, and nobody has validated them.
You are concerned about confidential data going into AI tools.
You want your team trained before AI use spreads further.
You probably need this if one of these sounds familiar
AI that supports the work, under control
What AI may be used for, by whom, and with which controls.
Validation of the AI tools you use, sized to their risk.
Training on using AI within the procedure.
I build AI agents for QA/RA work myself, so the controls come from practice.
A procedure for using AI
Validated tools
A trained team
Experience from daily use
From a first conversation to the work itself
You tell me about the device, the market you are aiming for and your timelines. I tell you what I would do first.
When needed, I go through your procedures, your technical file or your most recent audit reports.
The scope of work, estimated hours and a timeline. I work by the hour.
I work with your team on what the company needs at that stage, and we meet regularly for updates.
A first conversation
Looking at what you have
A clear proposal
The work itself
I use AI in my own work every day. It drafts procedures with me and walks me through CAPA forms: after about 20 minutes of conversation, a CAPA is almost ready. Nothing goes into a quality system before I check it against the requirements, because AI can sound sure of itself and still be wrong. I help companies set up the same discipline: clear rules for what AI may do, and a person who checks every output.
How I use AI myself
Yes, with control. I help you write the procedure for using AI, validate the tools you use and train your team, so the AI supports your work without creating new risks.
Yes. I work with software as a medical device and devices with software, including AI-enabled devices, under IEC 62304, IEC 82304-1, EU MDR and the EU AI Act. Cybersecurity is outside my scope.
Both are possible. Many clients start with a project and continue with a few hours a month of ongoing support. I work by the hour, and offer a fixed price only when the scope is closed and known in advance.
Can we use AI tools in our quality system?
Do you work with software and AI-based medical devices?
How do engagements typically work – project-based or ongoing?
Related articles
A 30-minute call. You tell me where you are, I tell you what I'd do first and whether I'm the right fit.