MS Money with AI?

[JB]

Hello.

Has anyone experimented giving access to claude or ChatGPT to control money?

I really want to try it but I cannot think of where to start, which connector or MCP or how to develop one.

Thanks!

[Stephen Google]

I've used CoPIlot to compare fund options which saves time. Otherwise trusting an Artifical Idiot is worse than trusting a money grabbing financial advisor!

[Harry Walden]

You can't give AI direct access to Microsoft Money data because the databases are not accessible. Depending on the AI, you may be able to give it application control to press buttons, click links and fill in fields, but I have not tried that. However, I recently uploaded the entire three decades of my financial data to Gemini to see what is possible. The good news is AI (or Gemini at least) can read QIF and CSV files, so getting the data into AI is easy. Gemini can also create QIF files, so you can get AI output back into Money. The bad news is that AI hallucinates, especially when doing bulk work, so your prompts need to be locked down tight, you need to check the outcomes, and you need to make backups and take care when importing.

So far it has done basic analysis (because right now that is all I have asked for), found errors that I made decades ago so I could fix them, and calculated missing record fields. I also got it to rewrite thousands of records in QIF format because I wanted to correct a logic problem I made decades ago when setting up how to track superannuation...but this required me to do a lot of checks. That took a long time, but certainly much less than if I had tried to manual recreate the records.

I have been thinking about doing a write up of my findings so far,  but in the meantime I'm happy to answer any questions people might have.

For those concerned about privacy, QIF files contain no private information, just data records. Those data records do create a total picture of wealth and shopping patterns, but that is nothing more than I see when Google, Meta and Amazon reflect back to me advertisements for websites I have been looking at - but as always, you do you!

[Stephen Google]

Let's hope those QIF files don't include your account numbers enabling the Artifical Idiot to raid your accounts!

[Harry Walden]

They don't! Have you never looked inside a QIF file? 🤷🏼‍♂️

[Tom Roberts]

Thanks Harry for your v8ews and generous offer.

These AI models are under constant revision and getting more powerful.  

I had a funny incident about 6 weeks ago when i got ai to translate some pages of a biography. I realized that i'd forgotten to include a couple pages in the middle but the model spewed out a completed translation. I asked the model how it was able to translate the sections I forgot to include and it said it made up a narrative uo match the missing sections ! The model has often appologized to me.

--
You received this message because you are subscribed to the Google Groups "Microsoft Money" group.
To unsubscribe from this group and stop receiving emails from it, send an email to microsoft-mon...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/microsoft-money/59522b41-aacc-42c3-8944-881f4526ad8an%40googlegroups.com.

[Jaime Babb]

Hello

 

Definitely a long way to go in AI, however, it has created a lot of value for me in several areas.

 

I have already automated many proceses using claude+n8n+excel but I have to import transactions into MS Money and I am thinking that maybe I could do it with AI+application control.  I have tried macros but it does not work for me.

 

Which one would you use?

 

Thanks

To view this discussion visit https://groups.google.com/d/msgid/microsoft-money/CAEDzrA9c-GzXRKi8TJChkLHGUo_uHw5fUyEtvTa4XzARA9K6cA%40mail.gmail.com.

[David Kimber]

I have not tried converting.mny files yet, but this is what Gemini says about the topic. https://g.co/gemini/share/581f86222b2c , so getting the data in an open schema would probably allow some additional data correlation.

Would be good to hear if anyone has tried this.

Regards David

To view this discussion visit https://groups.google.com/d/msgid/microsoft-money/019b01dcd8ff%24c7c0be70%2457423b50%24%40gmail.com.

[Harry Walden]

I don’t understand why you would want to do this? What would be the purpose of putting it into another database when Microsoft Money already is a relational database, and a pretty damn good one.

[David Kimber]

Converting to an open environment and hosting in the cloud would have advantages to use from any device at any time vs a dedicated local MS OS world.

Regards
David Kimber
Mobile +44 77 333 11 972

To view this discussion visit https://groups.google.com/d/msgid/microsoft-money/89e11c5e-1dbc-4d6e-a904-8e0fc3eca108n%40googlegroups.com.

[Matt Petersen]

I balance all of my accounts (checking, savings, credit cards, escrow) to my bank statements.

Yesterday, I was struggling to find a discrepancy, so I took a screen grab of my Money account balancing screen and uploaded it to Copilot, along with my statement and asked it to find the discrepancy.  Took about 30 seconds and the results were perfect.  This task of checking one by one manually against the statement to find a $4.12 missing transaction has, in the past, taken way more time than I care to admit.

One cool detail, is that even though the dates on the statement are different than the dates I have in Money, Copilot was totally cool with that.

[David Kimber]

Thanks for trying! Possibly worthwhile continuing the Gemini response and asking how to resolve (it gave me a few suggestions). Apologies, I can't help currently as I am travelling and don't have access to a MS OS and consequently MS Access. However, I do wonder about the difference between Actual Intelligence and Artificial Ignorance in cases like this.

Regards
David Kimber
Mobile +44 77 333 11 972

On Fri, 1 May 2026 at 18:32, Brian Pearce <brianp...@btinternet.com> wrote:

I have tried converting the file using the python script. I found that for the UK version at least that the first 4 bytes should not be changed - I think they are part of the file addressing function used by the operatng system.  In my case 0100

 

The data to be replaced is data[4,19].  This seemed to go OK but MS ACess wants a password - there was none on the original file.  I added one but that didn't work.

 

I suspect that there may be a password embedded in the file which MS Money knows and so can open the file.

 

I will poke around but if anyone has any information on this please share.

 

Regards

 

Brian

 

 

---

From: microso...@googlegroups.com [mailto:microso...@googlegroups.com] On Behalf Of David Kimber
Sent: 01 May 2026 06:02
To: Jaime Babb
Cc: Tom Roberts; Harry Walden; Microsoft Money

To view this discussion visit https://groups.google.com/d/msgid/microsoft-money/CAFX9XWfxBzD32PwDzmjYRTMNySrt8_m3%3D6TnqmZ19_wgTQdoGQ%40mail.gmail.com.

[Tom Roberts]

i, too, have spent too much time reconciling at times. thx for sharing your success

To view this discussion visit https://groups.google.com/d/msgid/microsoft-money/bab4f921-b113-4417-bcd6-9c11c57864c3n%40googlegroups.com.

[Harry Walden]

I can see that advantage, but surely the better way would be to simply move to a more modern finance management program that has all those features plus all the security you would expect? The reason I have stuck with Money for decades is that I can't find anything better that suits my workflow and built up knowledge...and trust me I have gone through that exercise many times. To me, an Access database would feel like a complete backwards step no matter how much openness it provided.

[David Kimber]

Agree Access is not the goal, understanding the data schema and then being able to move our legacy data to a more modern environment is a need of many of us. We love the functionality but not the platform. A good data schema is just step 1 though, but why not start with the existing Money schema.

Regards David

To view this discussion visit https://groups.google.com/d/msgid/microsoft-money/ddefd8c4-486e-4f26-a5dc-37c3a9497dden%40googlegroups.com.