I'm reasonably convinced by now that the tools can add significant value to an organisation, they solve problems that I've seen in multiple banks and lenders in the industry, and conversations with my network of colleagues in the industry have been positive. This industry is small in numbers, but the clients are banks and lenders, so high potential value.
There is still development work to be done, but looking ahead, I'm starting to think about a business model which could monetise these tools. So far, I've thought of a couple of different models, and would appreciate any thoughts or input regarding them or additional options.
1. Sell the software directly. I'm not sure this is the correct approach, as the tools are aimed at IC's in the space, to be used interactively and in production to build and deploy models (think of it as a toolkit to be used directly in Python). There are complexities with licensing, and the code would need to be source-open to the customer due to the nature of Python packages.
2. Open source the project and sell consulting work to integrate the tools into lenders. The main issue with this approach is the limited number of potential customers (estimated around 30). If even a handful use the software without engaging in other paid work, it's a significant chunk of the potential market.
What do you mean by complexity in licensing? Do you mean that you don't own all the IP for the software? In which case, yeah, you're likely to struggle. Similarly if some or all is GPL'd (or similar) and the usage model implies a derivative work, that's probably not going to fly either.
I can't see any benefit to you in open sourcing wrt selling to financial institutions. You absolutely can go that route but, if you're happy to sell directly, that's will require less work from the customer side.
Also, even if you're not finished yet, if you have contacts who are genuinely interested in the work, have you thought about asking them to fund the completion? We've done that in the past, and I worked for a startup that had their first client pay them to complete the offering. The nice thing about that approach is that they're often not proprietary about it. As in, they're happy for you to sell to their competitors as it spreads the risk around.
Just open sourcing it (under a liberal license) may invite freeloading and potentially forking. Regarding less liberal licensing, many companies will not even consider any (A)GPL code, so a restrictive license is likely not making your business more successful.
So what remains? If you can make a fantastic product that's way ahead of the market, the license and mode of delivery does not matter much, you can even keep it closed source as long as you provide the customers a continuity strategy (e.g. via a code escrow company that will open source it in case your company ceases to exist).
If your product isn't that much ahead of other components, open sourcing it may boost popularity but can not guarantee you customers or income.
Edit: typo, reworded second paragraph.
Is there any chance to build a web app on top of the library which people would use? This could be a way to monetize, although I don't see ICs using such a tool.
Another way could be consulting/service based on your library. Let's say the library produces results and you can sell these results as a report to the company. The boundaries between consulting and service in such cases are quite fluid in my opinion.
Anyway, as others have mentioned, selling libraries to ICs is hard. Most work on their own stuff that fits their needs. Coding is also a way to understand and to build up intuition about things, so not sure.
I would actually go talk to a few people you think you could sell to at some time and figure out what they think about using such a lib.
1. Sanitizer 2. intelligence Logic 3. Show it as html
Monetizing would be possible by giving access to the service. Of course, it's your knowledge and you see what data formats/data exchange is needed. That would be the culprit.
It's gonna be hard selling to a bunch of guys, nearly all of whom privately entertain spinning off their own analytics library written in Python.