Sla direct naar inhoud

Where your data lives, who has access and what AI does with it

A company brain holds your brand, your offer and the way you work. Here is how I work with that, and what this site itself measures and makes with AI.

Where the brain lives and who has access

In its own database in the EU. In my name during the build; after delivery and payment in your company's name.

  • I often build in Supabase, but it can also be another database your company wants to use. After the handover it is in your company's name.
  • The database is in an EU region of Supabase, for example Frankfurt. I record the region per client.
  • Skills fetch live data, such as from your CRM, only for one task. They do not store it in the brain and write nothing back: a proposal goes to a queue that a person approves.
  • I record this in the proposal and in a build log, in GitHub or in the documentation you want.
  • Your team has access, with the rights you give.
  • During the build everything is on my accounts. After the handover I only have access for the ongoing care, with your permission and a data processing agreement. If I stop, I am removed.
  • Every topic has an owner in your team. Only that owner changes it, as a new version.
  • AI makes suggestions and never changes the brain itself.
  • AI licences, such as Claude Team, you pay yourself.

What never goes in

The brain holds company knowledge, not files about people. I record this list in writing per client.

  • Customer data and call transcripts. A skill fetches them live from the source when a task needs it.
  • Personal data, with one exception: the name of a topic's owner, and of someone who gives a quote with written permission.
  • Special categories of personal data and HR files.
  • Bank and payment details, contract amounts and confidential financial information or acquisition plans.
  • Passwords, keys and other secrets.

Which AI services process your text

Your team works in a business Claude account in your company's name (Team or Enterprise).

  • Do you use the brain in another AI tool, through an MCP connection? Then that tool's privacy terms apply as well.
  • Claude does not only process your questions in the EU. Anthropic may route traffic to the US, Europe, Asia and Australia and stores business data in the US. Anthropic's data processing agreement and standard contractual clauses (SCCs) for the transfer to the US apply.
  • Anthropic does not train on business data by default. The exception is feedback you send yourself with thumbs up or down. You can turn that off in your account, and I recommend it.
  • I do not use your data to train models myself.
  • If everything must stay in the EU, we look at a route through an EU cloud region together.
  • The AI provider processes your data under its business terms; which provider and which agreements we record per client.

Agreements in every proposal

What gets recorded, what happens in an incident and what happens if you stop.

  • Only metadata: team, skill and time. No prompts and no answers.
  • That data is there to improve the brain, not to assess employees.
  • It is deleted after 90 days.
  • If your company has a works council, you arrange its consent yourself. I raise it in the intake and in the proposal.
  • If I notice an incident with your data, I report it to you no later than 24 hours after I notice it. You then report it to the Dutch Data Protection Authority within 72 hours if needed.
  • If you stop, you get everything in open formats, with no exit costs. Then I revoke my access, delete what I hold and confirm that in writing.
  • After payment I transfer the brain, the skills and the accounts to your company. From then on you can always keep working.

AI use on this site

This site is itself built with AI. This is what AI makes here, and where what you type goes.

  • I make blogs, translations and images with the help of AI, outside this site. I read and approve every article myself before it goes live.
  • Nothing publishes automatically: I publish myself.
  • The Prompt Crafter and the Lovable prompts are no longer on this site. What you entered there earlier has been deleted.
  • There is no chatbot on this site.

What this site measures itself

Without your choice I only count anonymously, without cookies and without an ID. Google Analytics and Leadinfo only load after your yes.

  • Always, also without consent: for each page viewed, the path, whether it was your first page, the domain of the site that sent you here, the UTM tags in the link, mobile or desktop (based on the screen width) and the language. A click on a contact button is counted with only the type and the page, and for blog posts I count how far people read. I also count which parts of a page came into view, how far people scroll and which page of this site you clicked through from. Each count stands alone: there is no ID that links them to each other or to you. I do not count visits from robots and automated browsers that I can recognise. No cookies, no ID and no IP address. I keep this for 12 months.
  • Did you arrive through a link with UTM tags? Then your browser remembers those tags while this tab is open, so they go along in the link to Calendly. That shows me how many meetings each source brings. There is no ID in it.
  • Google Analytics from Google only loads with your consent. Google then sets cookies and sees which pages you view, how you arrived and which contact buttons you click. Advertising features are off.
  • Leadinfo only loads with your consent. Leadinfo uses the IP address of your network to recognise which company you belong to and stores an ID in your browser.
  • I keep your choice for 12 months in your browser. You change it through Cookie settings at the bottom of every page. If you withdraw consent, I delete the cookies and stored data of Google Analytics and Leadinfo in your browser.
  • In the contact and training forms, and when you book a call, I ask "How do you know me?". You pick LinkedIn, Google or Bing, AI assistant, Through someone or Other (up to 200 characters). I store your answer with your enquiry, together with the page you landed on and the UTM tags of that visit. They come from the page itself; I store nothing extra for this in your browser.
  • Every contact enquiry, training request and Calendly booking becomes a lead in my admin, one lead per person (by email address). It holds your name, email, organisation (the one you filled in, or else the domain of your email; not for free mail such as Gmail), a status, a note from me and an amount. For each enquiry I keep the date, the type, your answer to "How do you know me?", the channel and the page you landed on.
  • If you request a Brain Experience, I keep your email address (with your lead), the website, your answer to "Where do you want to save time?", the language and the page you landed on. I only read the public pages of that website. I keep the request for 2 years and your private page until 90 days after our last contact.
  • I fetch Calendly bookings from Calendly regularly: name, email, date and source of the booking, and your answer to "How do you know me?" if you filled it in. The link to Calendly carries the channel you came through and the page paths, if the link had no UTM tags yet. They go to Calendly with the booking.
  • I pass training requests on to the trainer.
  • I keep leads for 2 years after the last contact. If you become a client, I keep your data for 7 years after the last contact. Training requests I keep for 2 years. Contact enquiries and Calendly data also for 2 years, unless you are a client. This is cleaned up automatically every night.
  • The card "Help my brain get smarter" only counts how often the card is shown and which choice you make (yes, no or custom). Like the other counts, each count stands alone: without an ID, without a cookie and without an IP address.
  • What you enter in the contact form goes into this site's database and reaches me by email. The enquiry gets a short assessment based on budget and timeline.

The full list of retention periods and your rights is in the privacy policy.

Privacy policy

Questions about your data?

Ask them in the intro call, with no sales pitch.