How to use AI to review your stock portfolio

AI can help organize a portfolio review before you grant trading access. Give it current holdings, quantities, cash and the date of the snapshot, plus the goal and time horizon for that money.

Use a supported read-only connection where available. If uploading a statement, remove account numbers, addresses and other details the review does not need. Check the provider’s data terms before sharing financial records.

Review these holdings as of the stated date. Calculate position weights including cash. Identify concentration by company, sector and asset class, and overlapping ETF holdings using dated issuer data. Separate verified facts from estimates. Ask for missing information. Do not place, modify or cancel orders.

Then supply your goal: for example, long-term investing or money needed for a planned purchase. The same holdings can have very different implications for those two uses.

• The arithmetic. Do values match the broker and weights total approximately 100%, allowing for rounding? Does the answer include cash and identify any excluded account?
• The overlap. A direct holding can also sit inside several ETFs. Ask for the fund issuer’s holdings date and distinguish known exposure from an estimate.
• The recommendation. Does it follow from your goal, or is the assistant offering generic trades without cost basis, tax context or a time horizon?

For example, a $1,000 direct position plus $400 of indirect exposure through a fund is $1,400 of exposure to that company. Counting only the direct position understates the concentration. FINRA’s risk guide explains why concentration matters.

In Agency, you can ask research questions on Home without creating an agent. Do not fund an agent if you only want analysis: a funded agent can trade at its reviews.

Use the review to identify questions you can verify. If you later want ongoing management, turn the goal into a specific investing plan.