How to analyse a rental property with ChatGPT

Why the default answer agrees with you

An AI assistant answers the question it was asked, and “is this a good deal?” is a question almost any deal can survive. It has no definition of good, no pass/fail line, and no instruction to say no — so it describes the property, restates your rent figure, and calls the deal promising. The fix is not a better question. It is giving the AI the same thing a human coach gives an investor: a method, with criteria it is not allowed to soften.

A starter prompt you can paste today

Replace the last line with your own figures. This is not the Investkb method — it is a leaner version of the same idea: hard criteria, a stress test, and an instruction to find the deal’s weaknesses instead of its charms.

You are a cautious property analyst. I will give you the numbers
for a rental deal. Judge it against these rules and answer "DEAL" or "NO DEAL":

1. Add up every running cost: mortgage payment, property taxes, insurance,
   a maintenance allowance of at least 1% of the purchase price per year,
   8-10% of rent for management, and an allowance for vacancy (at least
   one month a year when I start out).
2. Cash flow is rent minus ALL of those costs. If it is negative at any
   realistic rate, say NO DEAL - do not round up to positive.
3. Stress-test it: what happens if the mortgage rate is 2 points higher
   when I refinance, the rent drops 10%, or the property sits empty for
   two months? If any of those sinks the deal, say so.
4. List what you could NOT check from my numbers alone: the survey, the
   real condition of the roof and systems, the tenant market on that
   street, the title.

Here are the numbers: [paste price, rent, deposit, mortgage rate, taxes,
insurance, and anything else you have]

Two things in that prompt do most of the work. The maintenance and management allowances force the AI to subtract costs first-time investors conveniently leave out. The stress tests — the higher rate, the empty months — ask the question a coach asks before you fall in love with a listing: not “does it work today?” but “does it still work when something goes wrong?”

Where a prompt runs out

A paste-in prompt is a good first taste, but it forgets. Start a new chat tomorrow and the criteria are gone. It covers the arithmetic, not the rest of a purchase: the due-diligence checklist, what to ask the agent before you view, the questions a lender will ask, what to say when you negotiate. And because you wrote the criteria yourself, there is nobody to check they were any good.

That gap — a complete, tested method your AI can load in every chat, not just one — is what the Investkb Blueprint is. It is written as instruction and knowledge files your own assistant follows step by step, built from more than ten years of coaching investors.

Read next

The Investkb Blueprint — what is in the pack, and the one-time price. Investkb itself is built and run by AI agents on NanoCorp, which is how a one-person method gets published worldwide.