Here's a scenario- stop me if it sounds familiar: you have a member of senior leadership who is newly entranced with MACHINE LEARNING or AI or IBM WATSON or whatever catchy phrase a 3rd party has promised to deliver, or invented software to lol

deliver, OR your CEO has promised Wall Street that your company is currently undertaking.

Well, realistically, *that* part is fine.  That's what those people do.  What's not fine is that they try to employ these fancy miracle techs to do the impossible; buying into the "uNcOvEr HiDdEn InSiGhTs In YoUr DaTa" hype and trying to teach a machine to do something that a human can't be taught to do.  They want that big win because it's sexy and stamps "NEXT PROMOTION" all over their career trajectory.  Just one little problem.

See, in industry, our processes (and machines) are already efficient- those that weren't… aren't in business any more.  There are no magical JSON parameters that will inform you about a machine's likelihood to purchase something.  These are regression problems we face, and there are no big, Lombardi, pack-it-all-up-boys wins. 
 

What there is, however, is success by a thousand cuts.  There are significant G&A gains, increases in employee satisfaction, and process efficiencies out there to be gained by the implementation of these technologies- we just have to adjust the diopter on our scope.  Think of the things that take time and manpower to accomplish: the coding of invoices; emailing reports to partners; presentation materials for The Board; tax documentation input.  None of these are going to headline the "big win" category, but the sum of the parts will be of greater impact than chasing the mythical ROI beast.

In short- if your village is being overrun by goblins, don't go hunting dragons.

Fus Ro GAHHHHHHH