Bounded AI and How We Work

This week's work concerned the Build page and the language that surrounds it. We revised our positioning towards what we have come to call bounded AI — capability with clearly stated limits. The phrasing now lives in both the English and German dictionaries, so the promise reads the same in either language: we describe what the systems do, and we are plain about where they stop.

The Build page gained a "How we work" section that sets out the sequence explicitly: a Statement of Work first, then Apuna Care afterwards. This is a matter of order rather than ornament. A client should know what is being agreed before ongoing support enters the arrangement. The aquaristik case study was split out from its original tickets and is held pending both a proper specification and consent — it will not appear until those conditions are met.

On the housekeeping side, the contact form bug is closed now that the Brevo key is confirmed in place, and the backlog records an accepted trade-off around branch protection rather than quietly ignoring it. Small entries, but they hold the standard: a decision is written down, and a human signs it off. That is the architecture we prefer to leave behind us.