I was sailing with friend recently and we had time to talk in a long night of low wind. He was a police officer who got interested into computers in the early days. He had the idea that the policeman's notebook would be a great application of personal computers. He got a sponsor in the boss of the police chief and was given the opportunity to develop the system. He then was given a small team of policeman to test his system. With their help the system was refined and really helped these fellows with their paperwork. As they rolled this system out he trained the team to train and support the rest of the force. The system was a great success. He received several awards for this system as well as many research grants.
In my opinion the key to his success was first that he was a policeman so he understood the job and the language. Then he enlisted the help of ready police officers to produce some rapid results. Then he used the same guys to train and support the rest of the men. A great credit to data processing manager at the time he also supported the project. He did not feel threatened by the project. How good is that!
How many times do we take the time to really train people to use new systems and get their feedback on features. I found this a great example of assessing readiness and producing rapid results, then building on the results to create a great new product.