We mitigated product risk while building AI features (blog.dopt.com)

Building AI features is inherently risky — much riskier than traditional non-AI features. One of the reasons it’s so risky is there’s a chicken-and-egg problem: you can’t know the quality of an AI response without building the system to see an actual AI response.

I’m curious why this is an AI powered (generative?) feature instead of pre-generated responses? Why was AI the right architecture choice?