It is tempting to build systems that sort people quickly. A quiz, a score, a label — and then a recommendation. It feels efficient. It is also, most of the time, wrong.
The cost of premature categorisation
Once a system labels someone — "avoidant," "high-conflict," "at risk" — that label starts to shape everything downstream. Questions get narrower. Interpretation gets biased toward confirming the label. The person stops being read in context and starts being read through a category.
We've watched this happen in hiring tools, in risk-assessment software, in therapy-adjacent apps that mean well and still do harm.
What we do instead
Layers is built to ask before it concludes. Context is gathered first — what's actually happening, between whom, under what pressure — before any interpretation is offered. Categorisation, if it happens at all, comes last, and always provisionally.
This is slower. It is also the only way we know of to stay honest about how little a single interaction can tell you about a person or a relationship.
Curiosity first. Judgment, if needed, last — and always open to revision.