Prototyping a given
And it has a cost you might have to eat. We are an IBM i5 server shop for LoB systems, many of which are still old school green screen. I use Sencha Extjs to write HTML5 apps using i5 as the server.
We started making a new product which the tooling on the machines is a crucial component of success. Hundreds of parts, eventually thousands. So the company built a toolroom with defined locations to track tools going in and out. So early on, one of engineers (not a user actually making any transactions on tools) wanted to be able to use barcodes to scan the tools in and out of toolroom with a mobile device.
Absolutely nothing wrong with idea, although I encouraged them to walk before they ran. So wrote the core server code, with green screen interface, to accomplish the needed transactions. Keep in mind the biz requirements were not even fully developed yet. The engineer thought it would be like grocery store where you scan one code and are done. Reality was the parts showed up from vendors with no barcode you could even scan, you had to look up what our Part No was by using the vendor's SKU.
Then when they used part, they needed to know which machine was going to use the part. So that meant a drop down box to choose which machine. And you also had to specify which toolroom location you were taking from. I could "suggest" the default location from server once they told me which part. But we have overflow areas, they could come from more than one place. So that is another drop down you may have to use. Then you have qty, which defaults to one but may have to be changed.
So you get point, this simple "scan and go" app was not going to work like a grocery store, different requirements. To even prootype this for them, I had to buy a $300 iPad mini and barcode scanner peripheral for $500. In hindsight, should have prototyped with built in camera to read barcode. Too clunky for production use but would have worked for prototype. Because as you might have guessed, by time the actual users tested and realized they had to use all the drop down inputs and virtual keyboard (if qty not one) on a machine issue, plus have to look up our Part No from vendor SKU on put away of new part, they ended up just using the green screen terminal which was just sitting there in toolroom anyway.
So a perfectly good hardware of $800 with an app optimized for what it needed to do is just sitting there collecting dust. As a developer, I loved the experience. I had been waiting for opportunity to write a Touch app on a mobile device. So I've added that to my skillset for future. But it did not payoff from company's point of view and really nothing I can do about it unless they change the biz requirements of what the app needs to do.
But until actual shopfloor users could touch the device and use the app, they could not give any feedback. If you are writing phone app, no investment other than writing the code. But in our case, took a small investment just get something in front of users to give feedback on.