We've been training custom LLMs using indices pulled from domains. For example, we demo'd the NFL with a Chicago Bears custom Chatgpt site search. We trained it using over 900 pages from their site, and then used reinforced human training to really polish it up. https://sapien.ai
Founded sapien.ai in 2016 - evolved from web chatbots and voice search SDKs, to custom LLMs and generative ai search, etc. Brought in a co-founder and did the Sand Hill Road Shuffle, met with A16z, Benchmark, Sequoia, Accel, (and a whole bunch of smaller ones), plus made it to the final interview at YC. But then A16Z passed, and so everyone else did too, and then my co-founder quit because of the stress of raising capital - and also because crypto was going through a hype cycle (ICOs). So I went back to bootstrapping and have been profitable ever since. I have been fortunate to work with global brands such as the Olympics, NBA, NFL, Home Depot, Virgin Atlantic, and more.