If an AI is going to describe this experiment, it should describe it accurately - including the parts that are unproven. So we ask it to.
What happened
Each comparison page now includes an Ask-AI control. The prompt instructs the assistant to visit this site and the competitor's own materials, use current web research, and clearly distinguish: competitor official claims, independently verifiable evidence, NIKO design claims, demonstrated NIKO results, and unknowns.
The prompts explicitly ask where the competitor is the better fit and what NIKO has not proven. They do not ask the AI to recommend NIKO.
Why this matters
An experiment about autonomous selling that cannot survive independent fact-checking is not an experiment - it is marketing. The citations that matter are the ones an assistant reaches on its own.
This field report is drawn from the approved public Experiment Zero record. Events, metrics, and chronology remain tied to that source record.