I decided to revert to something I did earlier in my career: putting everything down on a spreadsheet. This time, I had a Master's in Accounting (a story for another day), so I set it up in accordance with Generally Accepted Accounting Principles (GAAP). This approach to accounting is meant to make it easier for business stakeholders to understand the business's state and trajectory, and I wanted those benefits on a personal level. I set up my accounts, I did my projections, and lo and behold, not only could I afford to go to Trinidad for Christmas, but I also saw clearly what I could do to improve my financial prospects.
Being able to lay out the numbers and plan my finances, rather than trying to do it all in my head, improved my financial well-being. Your financial well-being is how you feel about your financial position, regardless of how much money you make. How you feel about your ability to do the things you want to do and get the things you want to get is what makes up your financial well-being. Even financial experts and economic professors could be financially literate but financially unwell.
After using the spreadsheet for a few months, I realized that what I had built could be useful to other people on a similar journey to mine. We all want a better handle on our finances, but people in tech build a lot of the "budget apps" that exist with no knowledge of accounting, which is the language of finance. I started a project called "Finexity" that was meant to abstract the messiness of accounting and use accounting reports to make it easier to assess your financial position.
Old habits die hard, and I realized I was spending a lot of time building a tool I thought others would use, without any real evidence to back it up. One of the interesting shifts with AI becoming more accessible has been that I get to focus more on the product and what it's meant to do and whom it's meant to serve, rather than the act of building it. With extensive experience writing code, I had already built a robust framework and development harness centered on my open-source project, WeOS. I realized I could focus on solving problems that matter to me without worrying about building for others. I scolded myself for losing focus on my most important customer and created a mini-me, a digital twin meant to solve the problems I find most cognitively taxing.
Finance is the first major area, and I'm also working on bringing Food and other facets online. My modus operandi has been to use WeOS as the store for my AI, powered by temporal events, skills, and a knowledge graph, and to have Claude (my LLM provider of choice) interface with it using MCP tools. This has opened a whole new way of doing agile development for me: I don't immediately start adding interfaces and code; instead, I interact with the data through Claude, and when patterns emerge, I enhance the interface. This has been a great way to mitigate the challenges of SaaS: it does 80% of what you want (the other 20% you could use Claude to accomplish), while reducing reliance on AI (and the associated token spend) by creating deterministic solutions for well-traveled paths.
This is a personal project that I still think others can benefit from (there is a decent accounting engine behind it), and I'd like feedback on what I have so far and what's next. If you're interested, you can take a peek at a demo instance or join the list to get your own.