PLATFORM · RESPONSIBLE AI
AI does the first pass. Qualified assessors review every judgment, sign it off and stay accountable for the result.
Each requirement marked, with reasons.
Review each judgment and change it where needed.
The assessor owns the result.
Every decision traceable.
Every AI output and every human change is logged.
MODELS
We use large language models from Anthropic, OpenAI and Google, and choose the model to suit the task.
Customer data is never used to train, fine-tune or improve any model, and our providers are contractually bound to the same.
We contract with model providers on zero-retention terms where they offer them.
AI providers may process data in the US, under recognized transfer safeguards.
LIMITS
It doesn’t make final decisions. Assessors do.
Its outputs are advisory and probabilistic, not a substitute for human judgment.
It doesn’t make solely automated decisions about individuals.
Questions about how we use AI?