Articles

Free editorial articles on AI in investment, comparing robo-advisers and human advice.

These are the standing articles on this site. Each one takes a narrow question about automated and advised investing and answers it from the evidence rather than from a sales deck. New pieces are added as questions arrive through the inquiry form.

What a robo-adviser actually does to your money

A robo-adviser is not a single algorithm but a pipeline. You answer a risk questionnaire; the system maps your answers to one of a handful of model portfolios built from low-cost index funds or ETFs; it allocates your deposit, and on a schedule it rebalances back to the target weights. Some platforms add tax-loss harvesting on taxable accounts. The portfolio itself is usually a sensible, diversified, market-cap-weighted one — the engineering is in the automation and the fee, not in a secret stock-picking model.

What you give up is specificity. The questionnaire cannot see that you are about to fund a business, support a parent, or hold a concentrated stock position from a former employer. It optimises the portfolio for the answers you gave, not for the life you actually have.

What a human adviser adds beyond the portfolio

The persistent finding in the academic literature is that the value of a human adviser is concentrated in behaviour and planning, not in security selection. An adviser is most useful at the moments an algorithm is least useful: when markets fall sharply and the right move is to do nothing, when a life event changes which assets should be held and in which account, and when tax planning across several accounts changes the after-tax outcome more than any fund choice does.

That value is real, but it is also uneven. It depends heavily on the individual adviser, on whether they are paid by fee or by commission, and on whether their incentive is to keep you invested or to move you into products. Reading the adviser's disclosure documents is the only reliable way to tell.

Where the two models overlap

The honest version of both routes looks similar at the portfolio level: broadly diversified, low-cost, regularly rebalanced, and tilted to your time horizon. A good robo-adviser and a good fee-only adviser will often recommend holdings that are hard to tell apart. The disagreement is about who reads your situation, how often, and at what cost — not about whether diversification works.

Reading the disclosures each side must give you

In Taiwan, a licensed adviser and a licensed automated platform each carry regulatory disclosures: how they are paid, what conflicts they have, and what recourse you have. We walk through where to find these in a typical agreement and what each line is actually telling you. If a disclosure is missing or vague, that itself is information.

Cost, in context

Robo-adviser all-in fees in Taiwan commonly run from roughly NT$0 to a small percentage of assets under management, often below the cost of a traditional advisory relationship. Human advisory fees vary more widely and may be flat, hourly, or a percentage. We compare the structures without ranking them, because the right fee depends on what you are paying for — portfolio management, planning, or both.

Suggest an article topic if a question you have is not covered above.