Write down your investment process.

Define your investment decision framework

Write down what the money is for and how big a loss you could take.

• Say when you’ll need the money, and which accounts this covers.
• Keep two limits apart: the drop you could sit through without selling, and the loss you could afford. If you haven’t worked one out, leave it open.
• Add any view of the world you want your decisions to follow, such as where interest rates are heading.

Watch out: If AI fills in your goals or your risk tolerance, those are its guesses, not yours.

Find an allocation that fits

Check each mix you’re considering against what you wrote in section 01.

• List the mixes you’ll compare, and date every return or risk figure you rely on.
• Write down how often you’ll review the mix and bring it back to its targets. Leave it blank until you’ve chosen a rule yourself.

Watch out: Treat an allocation AI suggests as one more option to compare. Get qualified advice if you need it.

Turn your outlook into themes and theses

• Write each view in your own words, then add when you expect it to play out and what would change your mind.
• Say what the view could change: your allocation, the stocks you pick, or nothing.

Watch out: Write down the evidence and the position size separately from how sure you feel.

Research and challenge the view

• Date each source, and keep what it says apart from what AI told you.
• Write down another explanation for what you’re seeing, and when you’ll look again.

Watch out: Find the claim your thesis leans on most in its original source, and note anything the source doesn’t say.

Tailor how each view is expressed

• For each fund or stock you consider, write down why the view should show up in its returns.
• Decide whether to act now, wait, or drop the idea, and write down why.

Watch out: Having more choices than a fund doesn’t make any of them right. Look at liquidity, costs, and overlap with what you already own before you choose.

Set up and automate the work

• List the chores you repeat, such as pulling prices or reading new filings, and where you keep each result.
• For each tool, write down what it may do on its own and what needs your approval.

Watch out: Break one automation on purpose, with a missing or stale input, and make sure it stops instead of making up an answer.

Review and refine the system

• Log each decision separately, with its date, even a decision to change nothing.
• Set the date or event for the next review, and note whether each automation is still worth its upkeep.

Watch out: Six months from now, you’ll want to know which parts were evidence, which were AI’s suggestions, and which were your call.