Chapter 15: Building a Plan That Holds

We've toured the machines, the data, the speed, the tools, the research, the plumbing, the failures, the frauds, the regulators, the due-diligence questions, and the tax code. Now the only question that matters: what do you actually do with all of this?

The core-and-satellite answer

A sound financial plan has a core: diversified, low-cost, tax-aware, rebalanced, matched to your goals and timeline, the boring machinery from Chapter 6 that automates discipline. The core is where your future lives. It doesn't need AI trading. It needs you to fund it consistently and leave it alone.

Around the core, you may keep a satellite: a small, separately accounted, hard-capped allocation for things like AI trading tools, individual stock bets, or a DIY bot. "Small" means an amount whose complete loss changes nothing, not your retirement date, not your kid's education, not your sleep. Five percent is a common ceiling. Many planners would say less. The satellite is budgeted like entertainment or tuition, because that's what it is: the price of curiosity and education.

The order matters. Fund the core first. Automate it. Then, only then, fund the satellite from what's left. Anyone selling you AI trading who wants the satellite funded before the core is built is selling you risk you can't afford, whatever the backtest says.

Realistic expectations

What AI can actually do for an everyday investor in 2026:

• Compress research time from hours to minutes: screening, transcript analysis, filing summaries, portfolio Q&A. This is the real edge, and it's the least glamorous claim in the industry.
• Automate discipline: rebalancing, tax-loss harvesting, contribution schedules, through low-cost robo-advisors. The evidence for this improving outcomes is the strongest in the book.
• Generate ideas worth investigating. Screeners and signal tools as starting points, never conclusions.
• Teach you how markets work. Building and paper-trading a system is a superb education, priced as tuition.

What it can't do:

• Replicate institutional advantages, data exclusivity, colocation, compute, talent, at retail prices.
• Turn a small account into a large one reliably. Fee math (Chapter 13) and tax math (Chapter 14) eat small edges alive.
• Replace judgment, diversification, or a plan. No model knows your goals, your timeline, or what a 30% drawdown does to your life.
• Provide certainty. Anyone selling certainty is selling fraud or self-deception.

The human edge

After fifteen chapters about machines, here is what the machines can't do, the durable human edges:

Goals. No AI knows what your money is for. Retiring at 60, funding a business, leaving something for your kids. The plan starts from the life, not the portfolio. Machines optimize numbers. Humans decide which numbers matter.

Behavior. The biggest determinant of investor outcomes has never been strategy selection. It's behavior: not selling in panics, not chasing in manias, rebalancing when it feels wrong. The best financial AI ever built is the one that keeps you from trading stupidly, which is why the boring robo-advisor beats the exciting trading bot for almost everyone.

Skepticism. Every scam in Chapter 11 worked by bypassing skepticism with urgency, authority, or greed. The investors who kept their money shared one trait: they asked rude questions and waited for answers. Verification is a human discipline. No tool does it for you.

Accountability. A fiduciary adviser answers to you, legally. A bot answers to no one. When you delegate decisions to machines, delegate to machines whose operators owe you a duty, and keep the final authority where it belongs: with the person whose future is at stake.

The last word

"Should I be using AI for trading?" The question from the Introduction, and my answer after all of this:

Use AI to understand your investments better, the research tools, the screening, the analysis. Automate the discipline, the rebalancing, the harvesting, the contributions. Keep anything resembling trading small, separate, and honest about its costs. Verify everything. Trust audited numbers over screenshots, fiduciaries over vendors, and your plan over your impulses.

The new world of AI trading is real, and it's remarkable. But the oldest truths of investing survived every previous revolution, diversification, low costs, tax awareness, discipline, skepticism, and they'll survive this one too. The machines changed the game. They didn't change the rules of winning it.

Build the plan first. Let the machines serve the plan. Never the reverse.