Here's an example of a design project that had actual production components in stock, ready to ship in just over two weeks after the first design sketch, and saved a boatload of money for the company:

At BigToys, we had (they still have it) a simple rope climber, with one end attached to the ground, and the other attached to the structure about 10 feet up. To fill orders, we had been using rope assemblies out of a large shipment of them, ordered over two or three years before. Suddenly, there was a spate of failures, where the ropes slipped out of the top connection, causing the kids to fall. We knew that we had to react quickly, because there were a lot of these out in the field (they were included on about 80% of all of our play structures that we sold).

Our management sent out a letter of recall, and the product engineer and I sat down to discuss the failures, and our options. It soon became clear that the reason for the failures was an undersized connector (a "swage", a metallic sleeve crushed onto the rope by the supplier). It was too short, and couldn't be relied on to give enough friction to hold the rope securely. The engineer said that he wished that we could put a longer swage onto the rope. It occurred to me that we could clamp a steel or cast aluminum device over the rope to both replace the original connection to the support pipe, but also crush a few inches of rope in the bargain, using simple nuts and bolts to compress two halves together. I sketched up the basic geometry, and took it into AutoCAD Mechanical Desktop to refine it and create a 3D model. This took about 2 days, at the end of which I had two similar versions, one with large "teeth" that would bite into the rope when assembled, and another that had many more teeth, but much smaller ones. (The engineer and I had each proposed a tooth variation, mine was the larger-toothed one.) The model was "hermaphroditic", meaning that each half fit onto itself to form a whole assembly. I included a cavity above the teeth that could capture the original swage, and sized it to fit the correct swage, so it could be used in production after the emergency replacement issue was past.

I saved the models as .stl files, scaling them up by 3% to compensate for shrink once they were cast in aluminum, and e-mailed them to a supplier who did 3D printing in plaster. He sent the plaster models of each variation back to us the next day, and we took them to the foundry to be cast into "function models" for testing. In two days, we had the castings in hand, and we load tested them that night in our test shop, to the ASTM safety requirements. (My large-tooth design passed, while the other failed! Chalk one up for the designers!)

I made some minor revisions to the CAD model in the morning and sent it off to the 3D printer again to be copied twice for creation of the casting pattern, and the following day it was taken to the pattern maker, who mounted them to a match plate and added runners, etc. Around one week later, we had the first order of components in our warehouse, tested them again for ASTM safety compliance, and began shipping out repair packs to our customers. Our VP of Finance told me that the casting had saved the company about $350,000 in replacement costs, including ropes and field labor. It replaced the original connector immediately, and is still in production today, eight years later.