MY EXPERIENCE:
RFSpot was a startup hoping to upend the retail industry by handling inventory checks by robots that would roam the store, using cameras and RFID to check inventory on the shelves. Currently this is done by employees, who once a week or so go over the whole store checking manually. RFSpot's robots, by automating this process, would make it more accurate, both by eliminating clerical error and running daily instead of a few times per month. The hardware (robots) were those ipads-on-broomsticks used for remote meetings by some companies. The system basically worked, and we were running tests with large retailers like Tesco in Great Britain and Home Depot in the US, but the company ran out of money and, after stringing the employees along for several months, finally closed its doors. While in 10 years I believe all large stores will be automated in this way, in 2015 the sensor technology was not quite where it needed to be. Someone will make a lot of money providing this service, but it won't be RFSpot (and me).
My part in the company was mainly communications infrastructure: I wrote programs that would automatically upload the large datafiles to AWS, run analysis, and present the results. I also did much of the source maintenance, generating daily releases from the GIT source.