Identity verification API you are currently using and does proprietary AI really mean anything or is it just gimmick?
If you want to save time by avoiding this big block of text, you can just skip to the end where I have asked the question.
So we're evaluating identity verification API and keep running into the same claim across every vendor's website. "proprietary AI" "built-in-hous" "our own models". It's on literally every homepage and at this point it's starting to feel like "artisanal" on a coffee shop menu, a word that used to mean something and now means nothing. But here's the thing, I'm not sure its entirely meaningless either. Like there is probably a real difference between a vendor that actually owns and trains their own models versus one that is stitching together AWS recognition, a third party OCR library and a watchlist API and calling it a platform. The question is how do you actually tell the difference from the outside when every vendor is making the same claim. From an API integration standpoint i also care about the practical stuff. Latency at scale, sdks that don't feel like they were written in 2015, webhook reliability and whether the documentation is written for developers or for the sales team's Powerpoint. We're building on this so the developer experience matters as much as the accuracy numbers.
So what are you guys actually integ͏rating in production right now and has anyone done the work of figuring out which proprietary AI claims are true?