Comment on Any self hosted personal finance projects doing anything interesting with AI that you've found value in?

• chazwhiz@lemmy.world ⁨6⁩ ⁨months⁩ ago

  I found this which is overkill for personal use (plus not self hosted etc) but does a good job of laying out this sort of application: midday.ai/…/automatic-reconciliation-engine/

  “Instead of just comparing text strings, we use 768-dimensional vector embeddings to capture the semantic meaning of transactions and receipts.

  // Generate embeddings for transaction data
  const transactionText = prepareTransactionText({
    name: transaction.name,
    counterpartyName: transaction.counterpartyName,
    merchantName: transaction.merchantName,
    description: transaction.description
  });
  
  const embedding = await generateEmbeddings([transactionText]);
  

  These embeddings allow our system to understand that “AMZN MKTP” and “Amazon Marketplace Purchase” refer to the same thing, even though the text strings are completely different. The system learns patterns like:

  • “SQ *COFFEE SHOP” → “Square Coffee Shop Receipt”
  • “PAYPAL *DIGITALOCEAN” → “DigitalOcean Invoice via PayPal”
  • “APL*APPLE.COM” → “Apple App Store Purchase””

  source
• chazwhiz@lemmy.world ⁨6⁩ ⁨months⁩ ago

  Yep, that’s exactly the sort of thing I’m thinking about here. And it doesn’t even need to be full on chat style LLM, just some decent NLP that can recognize WALMART, WAL-MART, or WMART are all the same thing and label it.

  But for some reason this question brings out all the assumption people who want to give financial advice or talk about the AI image the saw last year with 6 fingers.

  source

  • Mordikan@kbin.earth ⁨6⁩ ⁨months⁩ ago

    Honestly, you don't even need NLP for this. Excel supports regex now so you could just do a call like =REGEXTEST(A1, "(?!)^w.*mart$"). Then just mark by type and graph out to see where your main spending is coming from.

    source

    • chazwhiz@lemmy.world ⁨6⁩ ⁨months⁩ ago

      You’re missing the point, that would require sitting down and manually doing that for every conceivable payee. Walmart is just an example. The value of any sort of “intelligent” component would be for this to happen automatically and seamlessly for the user. Hell, the AI layer could just be “write regex for al the possible similar payees across these documents”.

      source

      • Mordikan@kbin.earth ⁨6⁩ ⁨months⁩ ago

        that would require sitting down and manually doing that for every conceivable payee

        That's just called VLOOKUP(). I think you're over-complicating this process. If you sit down and look at your finances, you'll notice that the number of payees you have isn't some absurd unmanageable amount. As others have mentioned, there's no real use case for involving AI this way. There's no scale, no real benefit to financial tracking, etc. I get this is just to use AI for the sake of using AI, but that's not really a goal when writing financial software.

        source
  • BlameThePeacock@lemmy.ca ⁨6⁩ ⁨months⁩ ago

    You aren’t going to find a useful AI system for personal use in Finance, you simply don’t have the scale needed to benefit from it. You don’t make enough transactions per month that looking over it yourself is going to be any slower than reading the AI summary.

    The question as with most process optimization and data analysis is, what’s the actual result you’re hoping for? If you want it to be able to summarize WALMART, WAL-MART and WMART so you can see those numbers added together, you already know you spent a lot at Walmart. Whare are you going to actually do with that information?

    source