AI Portfolio Analysis: What a Good Portfolio Review Must Deliver

A review worthy of the name goes beyond the position list and looks inside the funds:

• Concentration risks: look-through at the single-stock, sector and country level. With a €100,000 portfolio and an effective 25% US tech exposure, €25,000 hangs on a single narrative — more in our article on portfolio concentration risk.
• ETF overlap: combining MSCI World and S&P 500 means holding many stocks twice — overlap analysis quantifies it.
• Factor exposure: how strongly is the portfolio tilted toward growth, value, size or momentum — deliberately or by accident?
• Risk score: volatility, drawdown potential and correlations, condensed into one interpretable figure.
• Tax check: are the German saver’s allowance (€1,000, €2,000 for married couples) and loss pots being used? Is the FIFO principle factored into planned partial sales?

An LLM can comment on a pasted position list — nothing more. It lacks the fund holdings data for a look-through, current metrics and consistency: the same question yields a different answer tomorrow, and language models regularly hallucinate numbers. A specialized AI analysis, by contrast, works on verified holdings data, computes overlaps position by position and delivers reproducible results.

Then there is privacy: pasting your portfolio into a US chatbot is not the same as entrusting it to a platform with a clear GDPR basis.

The key distinction: a robo-advisor manages your money and makes the decisions for you — for roughly 0.3–1% in annual fees on top of fund costs. An analysis tool like MoneyPeak examines your existing portfolio and leaves the decisions to you. For self-directed investors that is the decisive difference: you keep control over broker, products and tax timing while getting the analytical depth otherwise reserved for wealth management.

When the managed route can still make sense is covered in our robo-advisor comparison.