I work for a major bank as a contract software developer. I've been there a few months, and without exception this place has the worst software practices I've ever seen. The software my team makes has no formal testing, terrible code (not reusable, hard to read, etc), minimal documentation, no defined development process and an absolutely sickening amount of waste due to bureaucratic overhead.

Part of my contract is to maintain a group of thousands of very poorly written batch jobs. When one of the jobs fails (read: crashes), it's a developers job to look at the source, figure out what's wrong, fix it, and check it in. There is no quality assurance process or auditing of the results whatsoever. Once the developer says "it works" a manager signs off and it goes into production. What's disturbing is that these jobs essentially grab market data and put it into a third-party risk management system, which provides the bank with critical intelligence. I've discovered the disturbing truth that this has been happening since the 90s and nobody really has evidence the system is getting the correct data!

Without going into details, an issue arose on Friday that was so horrible I actually stormed out of the place. I was ready to quit, but I decided to just get out to calm my nerves and possibly go back Monday.

I've been reflecting today on how to handle this. I have realized that, in probably less than 6 months, I could (with 2 other developers) remake a large component of this system. The new system would provide them with, as primary benefits, a maintainable codebase less prone to error and a solid QA framework. To do it properly I would have to be outside the bank, the internal bureaucracy is just too much. And moreover, I think a bank is fundamentally not a place that can make good software.

This is my plan.

1. Write a report explaining in depth all the problems with their current system
2. Explain why their software practices fail and generate a tremendous amount of error and waste. Use this as the basis for claiming the project must be developed externally.
3. Write a high level development plan, including what resources I will require
4. Hand 1,2,3 to my manager, hopefully he passes it up the chain. Worst case he fires me, but this isn't so bad.
5. Convinced Executive decides to award my company a contract for the new system

I have 8 years experience as a software contractor and have delivered my share of successful software products, but all working in-house for small/medium sized companies. When I read this over, I think I have a dynamite plan. But since this is the first time doing something this bold so I have my doubts.

My question is, is this a good idea? If you think not, please spare no detail.