AI Prompt to Summarize Everything You Owe and How the Total Has Moved

What you need

Paste the snapshot template below. Only the currency line and the liabilities section matter for this prompt, so you can leave the assets out. Give each debt its type and interest rate. Where you don't know a rate, write "unknown" so the model leaves that debt out of the average and tells you.

Below the snapshot, add two lines:

• Total debt 12 months ago: [amount, or "don't know"]
• New borrowing in the last 12 months: [amounts and what for, or "none"]

The first line shows the direction of travel. An old loan statement or your own records will usually give a figure close enough. The second line separates repayment from new borrowing. Without it, a total that barely moved could hide a new car loan that cancelled out a year of mortgage payments.

Include the debts people tend to forget, such as an overdraft you dip into every month, a buy now, pay later plan or money owed to family. Enter a debt in another currency in your snapshot currency and note the exchange rate you used, since a currency move can change the total without any payment being made.

Here is my financial snapshot. Use only these numbers.

Currency: [e.g. EUR]
Date: [today's date]
Age: [optional]

ASSETS
[name] — [category: cash / stocks / ETFs / crypto / property / metals / other] — [value]
[name] — [category] — [value]
...

LIABILITIES
[name] — [type: mortgage / car loan / credit card / student loan / other] — [outstanding balance] — [interest rate if known]
...

CONTEXT (optional but improves the answer)
Monthly income after tax: [amount]
Monthly spending: [amount]
Amount I add to savings or investments each month: [amount]
Value of the same assets 12 months ago: [amount]

The prompt

Copy this exact text and paste it into ChatGPT, Claude or Gemini directly beneath your snapshot block.

You are a financial analyst who explains personal debt in plain language. Using
only the liabilities in my snapshot above, plus the lines below, summarize what
I owe.

Total debt 12 months ago: [amount, or "don't know"]
New borrowing in the last 12 months: [amounts and what for, or "none"]

1. State my total debt and list each debt with its balance, rate and share of
   the total.
2. Group my debts by type and say which type dominates.
3. Calculate the average interest rate across all my debts, weighted by balance.
   Show the working and leave out any debt without a rate.
4. Compare today's total with 12 months ago. Split the change into repayment and
   new borrowing, and say whether a currency move played a part.
5. Describe how a debt mix like mine usually looks for people at a similar
   stage, and where mine differs.
6. Name the goals people commonly set for a profile like this, and the usual
   pitfalls.

If a number is missing, ask me for it instead of estimating. This is for my own
understanding: no borrowing or repayment recommendations, no lenders or products.

What good output looks like

The answer should open with the total and a table showing each debt's balance, rate and share. Say your snapshot is in Czech crowns: a 3,200,000 CZK mortgage at 4.5%, a 250,000 CZK car loan at 6.9%, 40,000 CZK on a credit card at 20%, and 4,000 EUR borrowed from a relative at 0%, entered as 100,000 CZK at an example rate of 25 CZK to the euro. The total is 3,590,000 CZK, and the mortgage alone is 89.1% of it.

The weighted average rate needs its working on the page: 144,000 + 17,250 + 8,000 + 0 = 169,250 CZK of yearly interest on 3,590,000 CZK, about 4.7%. A good answer adds that a simple average of the four rates would give 7.85%, which overstates what you pay, because the expensive card balance is small.

For the year-on-year step, if you owed 3,720,000 CZK a year ago and borrowed nothing new, the total fell by 130,000 CZK, or 3.5%, all of it repayment. Because one debt is in euros, the model should say that part of the change could come from the exchange rate.

The comparison with typical profiles should name the features of your mix, here one large secured loan with small unsecured ones beside it, and say plainly where yours departs from that pattern.

What this can't tell you

A summary is only as complete as your list. The model cannot find a debt you left out, and it cannot check balances against your lenders' statements.

What the model calls a typical debt profile comes from its general training and is no substitute for current statistics for your country or age group. The year-on-year change depends on the figure you give for 12 months ago, so a guessed number gives a guessed trend. The prompt says nothing about whether your debt is affordable, because it does not look at income or assets. Models also get weighted averages wrong, so redo that step with a calculator. Rules on interest, fees and tax relief differ by country.

Doing this without the retyping

Sumio has this as the "Liabilities Summary" prompt, one of the 62 built into the app. It runs on your total liabilities, using the debts you already track there, and a debt in another currency is converted to your base currency at the current exchange rate. Your records are stored on your device. When you run the summary, the figures it needs are sent to an AI provider to write the answer. The AI assistant is part of Sumio Premium.