How we write ​

I build AI in my own businesses before I help anyone else with it. The guides here come from that work: what really works at a Belgian SME, and what does not.

First built for ourselves ​

The approach behind Coronis was not thought up at a whiteboard. It comes from the daily use of AI on concrete, recurring work in my own businesses. What holds up there, I write down. What does not work, falls away.

That bar (would I use this myself?) determines what appears here. No theory that sounds good, but ways of working that survive practice. Every question from a reader that exposes a gap feeds the next piece.

Our editorial style ​

Every guide follows the same rules, whether you read it in Dutch, French or English:

  1. Clear language, no consultant jargon. Abbreviations are written out on first mention. We call things by their name. At the same time, we do not explain the basics endlessly.
  2. Sources alongside, in the text. Figures come from a source you can check and are linked. A figure without a source does not appear.
  3. Belgian context, no general theory. Examples are about the Flemish SME and about the reality here. Not about an abstract business somewhere.
  4. The voice of practice. Written by someone who builds and deploys AI himself, not by someone who writes about it from the outside. If a recommendation is awkward in practice, we say so.
  5. No invented figures, no marketing theatre. No "10x more revenue", no "transform your business". A claim we cannot back up, we take out or explicitly mark as an estimate.

How we use AI ​

Writing guides in three languages and keeping them current: that is a lot of work. Without help it simply does not happen. The interesting question is what "help" means here.

AI writes along. It delivers a first draft, a translation, a summary. But it goes through the same rails each time: fixed rules, a source check, and a human who approves before anything appears. The rules and the human gate are the constant. AI is the variable that frees up time.

Yes, AI sometimes makes things up. That is why there is that check. The source is verified, the claim tested, and nothing goes out without a human putting a hand on it.

Sceptical? Good. That is the right attitude towards anything built with AI. Twenty years in IT and innovation taught me what "tested and validated" really means. Your real test is not reading this page, but using the work. Find an error and tell me.

Who is behind it ​

Tom Janssens. Twenty years in IT and innovation management, including Eurocontrol (the organisation behind European air traffic control) and Belgian SMEs. Someone who turns AI into working solutions himself every day, not someone who only talks about it.

How we choose the sources ​

The source list is not fixed. It is kept up to date on the basis of what I actually read, and of what customers and colleagues bring in.

  • Official sources where the subject calls for it.
  • Recognised professional sources and research, linked in the text.
  • Questions that come back twice in conversations: a sign of a gap in the guides.
  • Recurring themes from search behaviour and customer conversations.

If a source delivers a lasting signal, it stays. If that stops, it falls away. External claims are linked. Click through and check.

When "last revised" changes ​

The date at the top of a guide only changes if a human has really read and adjusted it again. Not on a rebuild. Not on a translation round. Not "because it has been a while". Substantive changes are summarised.

What we do not do ​

  1. No invented figures. A figure without a source is not there, or it is marked as an estimate.
  2. No subjects for which we have neither experience nor a source.
  3. No text cut from other sites, sold as our own work.
  4. No clickbait titles that the content does not live up to.
  5. No passing on without a check. AI output without a human on it does not appear, in any language.

Feedback ​

See an error, disagree with an interpretation, or miss a subject? Let us know. Real errors are corrected quickly.

Want to know how we use that same system of rules and AI to build and deliver? Read how we work.