Ok well the prototype employee scheduler was designed for a call center. They were spending 3 days building a schedule as it has to conform to Erlang C/A, which is a queuing model that predicts how many calls will be taken on an infinite time scale if you have x number of agents answering the phone. It's on an infinite time scale because it reaches a stable state over a long enough period but it is rounded out to 30-60 minute blocks generally. Erlang C will overload if the call rate is too high for the agents, Erlang A models abandonment so always reaches a stable state. They both use a poisson arrival process as it is one of the few probability functions that can be modeled at high speed. Anything more complicated and we have to simulate it with random runs. The function is non-linear so some non intuitive things happen which is why it's used over just eyeballing rough guesses.
So I built a prototype that did it in 30 secs (down to 2-3 secs now with a new library that includes some additions to try and schedule employees fairly), got a meet with the fairly largeish company. They were like wow, this is great come back with a better looking cloud based prototype which I did. But then nobody wants to discuss anything further, potentially because more than one person's job would be removed if it were to be implemented but also possibly because I'm potentially aspergers and I may have upset the guy does the schedules during the meet as he inherited a system he doesn't understand and he said some stuff that was false concerning Erlang C and I called him out on it. Not to make him look bad just so we were dealing firmly in facts, but he seemed to take it personally.
So yeah, I'm here to learn how to market better based on my experiences.