Chapter 6: The Robo-Advisors Grow Up

If the trading bots of Chapter 5 are power tools, robo-advisors are appliances. Quieter, less exciting, and far more likely to actually improve your financial life. This is the corner of AI investing with the best evidence behind it, and in 2026 it's changing fast.

What robos actually do

Strip away the branding and a robo-advisor does four boring, valuable things. It builds you a diversified portfolio matched to your goals and risk tolerance. It rebalances when markets drift. It harvests tax losses, selling losers to offset gains, then buying similar-but-not-identical securities. And it does all of this for a fraction of a human adviser's fee. The industry now manages over $1.4 trillion.

The boringness is the point. The biggest driver of investor returns isn't picking the right stock. It's behavior: not panic-selling, not chasing fads, rebalancing mechanically. Automating discipline is the highest-value automation in personal finance. Everything else in this book is commentary.

The 2026 landscape

Betterment charges 0.25% a year (or $5 a month on small balances without recurring deposits) and launched an AI-enabled account recommender in March 2026. Wealthfront, also 0.25% with daily tax-loss harvesting, went public in December 2025, and its first quarters as a public company revealed something telling. Roughly three-quarters of its revenue came from cash management, not advisory fees. The "robo-advisor" business turns out to be substantially a cash business with an investing feature.

Schwab Intelligent Portfolios charges no advisory fee at all (with a $5,000 minimum), but keeps a meaningful slice of your portfolio in cash, which is its own quiet cost in rising markets. Fidelity Go is free under $25,000 and 0.35% above, adding tax-loss harvesting in 2026. SoFi (0.25%) added an AI Coach assistant in May 2026 and integrated Composer's strategy builder the following month. Robinhood Strategies crossed $2 billion in assets in 2026.

There's also been a consolidation wave worth watching. UBS wound down its Advice Advantage, U.S. Bank shut its Automated Investor, Schwab retired Intelligent Portfolios Premium. The industry is splitting along wealth lines, free or cheap automation for the mass market and human advisers for complexity, with the middle getting squeezed.

The new breed: AI-native fiduciary advice

October 2026 brought the most interesting launch in this space: M1 Advisor, an SEC-registered investment adviser offering AI-driven financial advice under the same fiduciary duties of care and loyalty that bind a human adviser, with regulatory filings to prove it (about $1.48 billion in regulatory assets across 11,400 accounts at launch). It's non-discretionary, meaning the AI advises and the client decides, and it launched free through the end of 2027.

Why does the fiduciary part matter so much? Because almost nothing else in this book owes you that duty. A signal vendor owes you nothing. A trading bot owes you nothing. An AI assistant inside a brokerage app owes you nothing unless it's structured as a registered adviser. M1's model, AI advice inside a fiduciary wrapper with the human making the final call, is the template worth watching in my view. Advice without accountability is just content.

Other entrants are circling the same idea from different angles: YC-backed Astor, Era (whose AI watches for overtrading and looming tax bills, guardrails rather than just gas pedals), and Range, which has openly discussed replacing human advisers. The direction is clear. The quality bar is not yet settled.

The agentic question, again

This is where I'd urge the most care. The industry is moving from advisory AI ("here's what I'd do") to agentic AI ("I've done it"). An agent that rebalances your portfolio is automating discipline. Good. An agent that day-trades your account, switches strategies midstream, and sizes positions on its own is automating action, and as one CFP memorably put it, "that's not automating discipline; that's automating action. Not the same thing."

Advisers are already warning about the specific failure mode: agentic platforms can generate enormous short-term capital-gains taxes that wipe out any extra return, and many of these platforms carry no fiduciary duty whatsoever. The tax chapter (Chapter 14) will show you the math. For now, the rule stands: the more authority you grant an AI over your money, the more carefully you should ask who it legally serves. If the answer isn't "you, under a fiduciary standard," the authority should stay with you.

Where robos fit in a real plan

For most investors, most of the time, the right relationship with AI investing looks like this. A low-cost robo-advisor or target-date-style automation handles the core portfolio, the diversified, rebalanced, tax-aware foundation, while anything resembling "AI trading" stays in a small, budgeted satellite role, if it exists at all. Chapter 15 will make this allocation explicit.

The robos won because they automated the boring stuff everyone needs. The lesson for the entire AI trading industry is right there: the most valuable financial AI isn't the one that trades brilliantly. It's the one that keeps you from trading stupidly.