How to ask good money questions

Most disappointing AI answers are caused by the question, not the model. This lesson is about asking better ones.

Why vague questions produce vague answers

Ask “how should I invest my money?” and you will get a generic paragraph about diversification and risk tolerance. It is not wrong. It is just useless, because the question could have been asked by anyone on earth.

The model has no idea whether you are 24 or 61, whether you have debt, what currency you earn in, or what you are saving for. Lacking that, it answers the average case — and nobody is the average case.

The four ingredients

A good financial question contains:

Context — your situation, in numbers, without identifying details
The actual decision — what you are choosing between
Your constraints — timeline, risk appetite, anything fixed
The output you want — a comparison, a list, an explanation, a calculation

You rarely need all four. Two or three transforms most questions.

Before and after

Retirement

Weak: How much do I need to retire?

Better: I am 34, spend about 3,200 a month, and want the option to stop working at 55. Assuming 4% withdrawal and 2.5% inflation, roughly what portfolio would I need, and what monthly contribution gets me there from 180,000 today? Show the assumptions you used.

The second gets you a number, the working behind it, and levers you can change.

Debt

Weak: Should I pay off my loan?

Better: I have a 12,000 car loan at 9% and 20,000 in savings earning 3%. I want to keep at least four months of expenses, which is 9,000. Compare paying the loan off in full versus keeping the cash, including what I would give up in flexibility.

Now the model has the actual trade-off rather than guessing at it.

Portfolio review

Weak: Is my portfolio good?

Better: My holdings are roughly 45% global index funds, 25% single technology shares, 20% bonds, 10% cash. I am 41 and would not need this money for at least 15 years. What risks stand out, and which of them would you address first?

Asking which to address first forces prioritisation instead of a list of generic observations.

Prompts worth keeping

Explain to a beginner:

Explain [term] in plain language, then give one example with real numbers, then tell me the most common mistake people make with it.

Force the other side:

Argue against [decision I am leaning towards]. What would have to be true for this to be a mistake?

This is the single most valuable financial prompt. AI assistants tend to agree with you; asking for the opposing case corrects the bias.

Surface what you forgot:

Here is my situation: [context]. What important questions am I not asking?

Make it concrete:

Give me the three things I should do this month, in order, with the reason for each.

Follow-ups matter more than the first question

Treat it as a conversation, not a search engine.

“Where does that number come from?”
“What if returns are 4% instead of 7%?”
“You assumed I have an emergency fund — I do not. What changes?”
“Explain that last part again, more simply.”

The second and third questions are usually where the value appears.

Three habits to keep

Ask for assumptions. Any projection rests on numbers the model chose. Make it show them and you can judge whether they are reasonable.

Verify anything jurisdiction-specific. Tax rules, contribution limits, pension treatment. Use the AI to understand the concept, then confirm the number with an official source.

Notice flattery. If every answer agrees with you, you are not getting analysis. Ask the opposing case explicitly.

The remaining limitation

Even a perfectly crafted question has a ceiling: you are typing your situation in by hand. It is a summary, it is approximate, and it is out of date the moment your portfolio moves.

The next step is letting the AI see the real numbers instead of your description of them. That is what the rest of this course builds towards.

Practice step

Take a money question you have genuinely wondered about, and write it twice — once the way you would normally, then again with context, constraints and the output you want.

Ask both. The gap between the two answers is the value of this lesson.

Module 2 complete. You know what AI is good and bad at, what to keep private, and how to ask.

Next: Module 3 — Your first portfolio, where we get your actual numbers into one place.