How to analyze your portfolio with AI

Most people know what they own. Far fewer can answer the second-order questions: is this portfolio what I think it is? Where's the concentration risk? Which theme is actually driving returns? AI tools like Claude and ChatGPT, connected to your real brokerage through Walnut, turn portfolio analysis into a plain-English conversation.

What portfolio analysis really means

Portfolio analysis isn't just “am I up or down?”, that's a glance, not analysis. Useful portfolio analysis answers:

Concentration. What percent is in any single name, sector, theme?
Attribution. Which positions drove gains? Which dragged?
Exposure vs intent. If you said “I'm investing in AI infrastructure,” does the portfolio actually reflect that?
Cash and liquidity. How much dry powder do you have, and how fast could you raise more?
Risks you haven't named. Currency, interest-rate, single-customer, regulatory.

Each of these is a question your broker's default dashboard rarely answers cleanly. AI plus Walnut can answer all of them.

How to analyze your portfolio with AI in Walnut

1. Connect your brokerage to Walnut

Sign up for Walnut and link your broker through SnapTrade. Your positions, balances, and recent performance sync automatically.

2. Connect Claude or ChatGPT

Generate an API key in Walnut's Connections page and paste it into your AI tool. Now Claude or ChatGPT can see your real portfolio when you ask it questions.

3. Start with the four questions

Run these four prompts in sequence, they take less than five minutes and surface most of what matters:

“Summarize my portfolio: total value, top 5 holdings by weight, top 5 by absolute gain or loss.”
“What's my concentration risk? Any single position over 10% of the portfolio?”
“Which of my baskets is performing best and worst over the last 30 days?”
“If I had to sell 20% of the portfolio to raise cash, what would you suggest selling, and why?”

4. Pressure-test the bear case for each big position

Pick your top three positions by weight. For each, ask the AI: “What's the strongest bear case here, and what would make me sell?” Write the answer down. If something the AI flagged comes true later, you'll know to act faster.

5. Schedule a recurring check-in

Block 10 minutes every Friday. Open the AI tool and ask: “What changed in my portfolio this week, and is any of it worth acting on?” Most weeks the answer is no, and that's fine.

Example: a real conversation

Say you own a 15-stock portfolio you've built around three themes, AI infrastructure, aging demographics, and clean energy. You connect Walnut to Claude and ask:

“Break down my portfolio by theme and show me which one is contributing most to year-to-date returns.”

The AI reads your basket assignments via Walnut, pulls each position's YTD return, weights it, and gives you an attribution table. You discover that AI infrastructure is 60% of the portfolio but 85% of gains, meaning a single theme is doing most of the work. That's an actionable insight you can't get from a broker dashboard.

Try it in Walnut

Walnut connects your brokerage to your AI tool so portfolio analysis is one conversation away.

Create a free Walnut account

FAQ

Why analyze my portfolio with AI instead of using my broker's tools?

Most brokers show what you own and basic returns. AI plus Walnut answers the second-order questions: are my positions concentrated, do my theme exposures match my intent, did one position drive most of my gains, what would I sell if I had to raise cash. Brokers don't help with that.

Can AI see my real portfolio?

Only when you connect it. Walnut acts as the secure bridge: you link your broker via SnapTrade and generate an API key for your AI tool. After that, Claude or ChatGPT can read your real positions when you ask portfolio questions, with row-level security keeping data scoped to you.

How often should I analyze my portfolio?

A monthly deep review and a weekly check-in is plenty for most investors. AI makes both fast, a monthly review that used to take an hour becomes a 10-minute conversation. Daily portfolio analysis usually correlates with worse outcomes, not better.

Does AI account for tax considerations?

AI can highlight large unrealized gains or losses in a position if your broker reports cost basis. But it isn't a tax advisor. Before harvesting losses or rebalancing across taxable and retirement accounts, consult a tax professional.