← Notes

The preference is not for a human; it is for not being stuck

What usability research records as well received is the bot owning the failure and offering an exit — which moves the question from "human or AI" to "solved or not solved".

The question that haunts anyone putting automated service in front of a customer is usually framed as preference: the person will prefer to talk to a human.

What the Nielsen Norman Group usability research records is narrower and more useful: "Owning the failure and offering an escape hatch (phone number or a live agent) were generally perceived favorably". The favourable rating is tied to a specific behaviour — acknowledging the failure and opening the exit — and not to a general preference for human service.

The difference changes what should be designed. If the preference were for a human, the conclusion would be to put people up front as early as possible, and automation would always be the lesser evil. If what is rated well is not being stuck, the conclusion is different: an agent that resolves quickly and acknowledges what it cannot resolve can be rated as well as human service.

It also suggests what to measure. Not satisfaction with "talking to a robot", which measures declared prejudice, but the outcome: whether the person got what they came for, in how long, and whether they found an exit when they asked for one. Those are questions the recorded conversation already answers, with no opinion survey.