Most SaaS companies are not doing anything particularly innovative or novel. Most of them provide value via putting in the work to glue together several other APIs, automating something that previously was harder to script/automate, or simply applying an idea that another company pioneered to another market segment: "It's like Theranos but for barbers!"
These use cases are generally so typical of the technology being used, that LLMs can do a lot of work to script things and it's usually pretty easy to QC.
I'm using LLMs to analyze 13F corporate filings, options flow, market news, etc. This was very labor intensive before. Essentially impossible for small companies. A 13F report can have 150 pages or so.
Just eliminating redious things. It's nothing difficult for a human, but we would have to hire a ton of humans and couldn't compete against the big guys then.
I'm not offering an LLM wrapper SaaS for my clients. We are just using this internally.
