What Is a Robo-Advisor and Should You Use One?

Sign up to a robo-advisor and the whole thing takes about ten minutes. You answer maybe fifteen questions: your age, roughly what you earn, what the money is for, how you would feel if the balance dropped by a fifth. A dial lands somewhere on a scale from cautious to adventurous. You are shown a pie chart, you press a button, and from that moment a diversified portfolio buys and rebalances itself without you ever logging in again.

That experience is genuinely useful. It is also worth understanding exactly what you just bought, because the marketing word "robo" does a lot of work here, and most of what it implies is not true.

What is actually happening behind the app

There is no artificial intelligence picking stocks. A robo-advisor does four mechanical things:

1. Sorts you. Your questionnaire answers map you onto one of a small number of pre-built model portfolios, often five to ten of them, running from mostly bonds to almost entirely equities.
2. Buys the model. Your money is spread across a handful of low-cost index funds or ETFs that make up that model. If you want the underlying logic, it is the passive side of the active versus passive argument, applied by default.
3. Rebalances. When shares run ahead and your 60/40 split drifts to 70/30, the system sells some of the winner and tops up the laggard to bring you back to target.
4. Handles the admin. Contributions, dividend reinvestment, tax reporting and, in some markets, tax-loss harvesting all happen without you doing anything.

That is the product. It is a questionnaire, a model portfolio and a rebalancing schedule. Whether it is worth its fee depends entirely on what you would have done otherwise.

The fee is three layers deep, not one

The advertised rate is the platform fee. It is not the whole cost, and this is where most comparisons go wrong.

Layer one: the platform or management fee. The headline percentage, charged annually on your balance.

Layer two: the fund charges. Every ETF or index fund the robo-advisor buys has its own ongoing charge, deducted inside the fund before the price you see. It never appears on your statement, so people forget it exists.

Layer three: trading and spread costs. Usually small for large mainstream funds, but not zero, and higher if the portfolio holds niche holdings that trade thinly.

The SEC makes exactly this point in its robo-adviser bulletin: even where the advisory fee looks low, investors should look separately at the costs of the ETFs held in the portfolio and any brokerage charges, because fees and other costs can significantly affect returns.

A worked example over 20 years

Take a 50,000 pot, in whichever of the three currencies applies to you, left alone for 20 years, and assume the underlying markets return 6 percent a year gross in every case. The only thing that differs is what gets skimmed off annually.

Route	Total annual cost	Value after 20 years
DIY: two index funds you buy yourself	0.15%	about 155,900
Robo-advisor: 0.25% platform plus 0.15% funds	0.40%	about 148,700
Human adviser: 1.00% advice plus 0.15% funds	1.15%	about 128,900

Those percentages are illustrative rather than quoted from any provider, and real returns will not be a smooth 6 percent. But the shape of the answer is the useful part. The robo-advisor costs roughly 7,000 more than doing it yourself across two decades. The human adviser costs roughly 27,000 more. The gap between the robot and the human is far larger than the gap between the robot and you.

So the honest question is not "is a robo-advisor cheap". It is "is the automation worth about 7,000 to me, and is human advice worth about 27,000".

Where robo-advisors genuinely earn it

They remove the decision. The most expensive investing mistake is not a bad fund. It is money sitting in cash for three years because you never felt informed enough to press go. A robo-advisor converts an intimidating decision into a form.

They rebalance when you would not. Rebalancing means selling the thing that has done well and buying the thing that has done badly, which feels wrong every single time. Automation does it without emotional negotiation.

They make regular investing invisible. A standing order into a model portfolio is dollar cost averaging with no willpower required.

They give you one number. Consolidated reporting across a whole portfolio is genuinely more useful than logging into three fund pages and doing sums.

Where they fall short

The SEC's own bulletin is unusually direct about the limits, and it is worth reading before you commit. Its warnings boil down to three:

The advice is limited to what the form asked. A robo-advisor may never ask about your credit card debt, your student loans, the property you own or the pension sitting with a former employer. It will happily invest money you would be better off using to clear a card charging 22 percent. The algorithm is not wrong; it simply never saw that part of your life.

Your answers might not match your behaviour. A questionnaire measures how you say you would react to a 20 percent fall. Nobody knows how they actually react until it happens.

Incentives are not always neutral. The SEC tells investors to ask whether the service is paid to offer particular products, whether it only offers funds affiliated with the parent company, and whether it receives referral or marketing fees. Ask that question in writing before you sign up.

The UK regulator found the same pattern from the supervisory side. In its multi-firm review of automated investment services, the FCA reported that most firms in its sample could not show they held adequate and up-to-date information about clients during an ongoing service, that disclosures were often unclear, and that some services relied on customers to self-identify as vulnerable rather than actively spotting them.

The country differences almost nobody mentions

This is where generic robo-advisor articles fall over, because the product is meaningfully different in each market.

United States. Tax-loss harvesting is the standard premium feature: the system sells a holding at a loss to offset gains elsewhere, then buys something similar. It only works within the wash sale rules in IRS Publication 550, which disallow a loss where you buy substantially identical stock or securities within 30 days either side of the sale. Two practical consequences: the replacement fund must not be substantially identical, and the rule follows you across accounts, so a purchase in your IRA can spoil a harvest done in your taxable brokerage account. Also check what the service actually covers, because most robo-advisors manage taxable accounts and IRAs but not the 401(k) sitting with your employer.

United Kingdom. The wrapper matters more than the algorithm. Inside a stocks and shares ISA, up to the 20,000 annual allowance for the 2026 to 2027 tax year, gains and income are not taxed, which means tax-loss harvesting has nothing to harvest. Any provider selling you tax optimisation on ISA money is selling you nothing. Check instead whether you are getting a discretionary managed service or an execution-only service with guidance attached, because the FCA specifically criticised firms comparing those two on cost alone without explaining the difference. Check the firm on the Financial Services Register and confirm your money is covered by the Financial Services Compensation Scheme if the firm fails.

Canada. The term is close to a misnomer. As the Ontario Securities Commission explains, Ontario's online investment advisers provide discretionary portfolio management with a registered adviser responsible for the decisions, so a human reviews your questionnaire, chases up inconsistent answers and signs off on the trades. That is a stronger consumer protection than a fully automated US-style service, and it is also why account minimums and service levels vary more than the marketing suggests. On tax, Canada's equivalent of the wash sale rule is the superficial loss rule: a loss is denied if you or an affiliated person buy identical property in the window from 30 days before to 30 days after the sale and still hold it 30 days later. The denied loss is added to the adjusted cost base of the replacement instead of vanishing. And as in the UK, harvesting is pointless inside a TFSA.

So should you use one?

A short decision, honestly answered.

Use one if: you have a lump sum or a monthly amount sitting in cash because investing feels like a project you keep postponing; you know you would not rebalance; your situation is straightforward, meaning employment income, no business, no property portfolio, no cross-border complications; and the alternative is doing nothing.

Skip it if: you would genuinely buy two or three broad index funds and leave them alone, in which case you are paying for a service you would perform for free; or your finances have moving parts an algorithm never asks about, in which case a few hours of paid, fee-only human advice will be worth far more than a percentage skim forever.

And in either case: clear expensive debt first, take any employer pension or 401(k) match first, and hold an emergency fund first. No portfolio, automated or otherwise, outperforms a 22 percent credit card.

The bottom line

A robo-advisor is not clever, and that is the point. It is a low-cost, disciplined way of doing the boring things that actually drive long-term returns: getting invested, staying diversified, rebalancing, and not touching it. Judge it on the total of the platform fee plus the underlying fund charges, verify the firm on your national register, and be clear-eyed that its advice can only reflect the questions it thought to ask. If that trade is worth roughly a third of a percent a year to you, use one. If you would happily do the same job yourself, keep the money.