Chapter 5: The Retail Toolbox

Now we enter World Two: the products sold to everyday investors with "AI" on the box. Some are genuinely useful. All of them are useful in narrower ways than their marketing suggests. This chapter is the straight tour, what each category does, what it costs, and what it realistically delivers.

A note on prices: every figure below was current as of roughly September–October 2026, drawn from public sources. Vendors change tiers constantly. Several did during 2026 alone. Treat these as ballpark, and check the vendor's site before spending anything.

AI screeners and signal platforms

Trade Ideas and its "Holly" AI is the heavyweight of this category. It scans more than 8,000 US stocks in real time. Overnight, the Holly engine runs millions of backtests across 60–70+ strategies and delivers a set of pre-market trade ideas, each with suggested entry, stop-loss, and target prices. The higher tier (Brokerage Plus) can auto-trade the signals. Pricing runs roughly $127 a month for Standard and $178–198 for Premium, with a free delayed-data tier. It's built for active US day traders, and it's genuinely powerful. It's also US-equities-only, the interface shows its age, there's a steep learning curve, and Holly's published win rates are backtest-derived. Live trading includes slippage the backtests don't.

TrendSpider (roughly $52 to $321 a month depending on tier) does automated technical analysis with natural-language scanning, so you can type what you're looking for in plain English, plus no-code backtesting and trainable custom AI indicators. It's probably the best all-in-one for a systematic active trader. The catch, common to the category: the most capable AI modules live in the upper pricing tiers, and the tool still requires your judgment to interpret what it finds.

Tickeron offers AI pattern recognition, intraday buy/sell signals, and predictive analytics across asset classes, priced per module. Danelfin takes a different approach. Instead of intraday signals, it scores US and European stocks and ETFs from 1 to 10 on expected three-month outperformance, built from thousands of fundamental, technical, and sentiment inputs. It's aimed at position traders, not day traders, but its methodology is opaque, so you're trusting a black box on a quarterly horizon. LevelFields is event-driven: it tracks two dozen catalyst types (FDA approvals, activist filings, buybacks) against five years of historical market reactions and assigns probabilities. Broad coverage, less transparent scoring.

The summary: these tools are idea generators, not money machines. A good screener compresses hours of manual chart review into minutes. None of them repeals the arithmetic of trading costs, which brings us to the most underappreciated number in retail AI trading.

The cost iceberg

Subscriptions are the visible part. Below the waterline: commissions, exchange and regulatory fees, market-data subscriptions, bid-ask spreads on every trade, slippage (the difference between the signal price and your actual fill), cloud or VPS hosting to keep bots running, and API/data fees.

One analysis making the rounds in 2026 put it starkly: a scalping bot showing 0.08% gross profit per trade can net something like 0.01% after all costs. For high-frequency-style strategies, infrastructure costs matter more than the subscription. This is why so many retail bots look brilliant in backtests and mediocre in live accounts. The backtest usually undercounts the iceberg. Chapter 9 will show you how professionals measure execution costs; Chapter 13 will show you how to demand the same accounting from vendors.

No-code bots

StockHero (around $50 a month) lets you run predefined or customized bots that execute automatically, no coding required, with a marketplace where you can rent other people's bots. "Hands-off" still needs supervision. Strategies decay, markets change regimes, and rented bots stack fees. Broker connections are limited compared to professional platforms.

Composer (roughly $10–40 a month across tiers) took a different tack: no-code portfolio rotation and AI-assisted strategy building for ETFs rather than day-trading signals. In June 2026, SoFi integrated Composer's strategy-builder technology, a sign of where the industry is consolidating. Whether the standalone product survives long-term was unclear at the time of writing. Another reminder to check current status before committing.

OptionAlpha offers no-code automated bots for options, including volatility forecasts and automated spreads. Options deserve a special warning: leverage magnifies edges and blowups symmetrically. An AI that is slightly wrong about direction with stock loses a little. The same error in options can lose everything. Automating options without deeply understanding options is how accounts go to zero quickly.

The 2026 shift: agentic brokerage AI

The newest development, and the one to watch, is agentic AI inside the brokerages themselves. Interactive Brokers lets users connect AI systems to portfolio data to generate trade instructions. Robinhood opened its brokerage infrastructure to external AI agents. eToro built its experience around "Tori," an AI assistant. Moomoo is opening infrastructure to personal AI agents.

The question to ask about each of these is not "does it use AI" but "how much authority does the AI get?" An assistant that explains your portfolio is a research tool. An agent that places trades, switches strategies, and sizes positions on its own is something else. It automates action, not discipline. Remember the CFP's line from Chapter 1. We'll return to it in the due-diligence chapters, because authority without accountability is where the real risk lives.

Copy trading and social AI

eToro, the social-trading giant, deserves its own mention. Its patented CopyTrader lets you automatically mirror up to 100 other traders' positions. In May 2025 it launched Alpha Portfolios: quantitative strategies trained on eToro's dataset of 40 million users' trading behavior, with a $10,000 minimum. The AI layer ranks traders and constructs portfolios from crowd behavior.

The limitations are structural. Past copier returns reflect survivorship. Traders who blew up disappear from the leaderboards. Herding risk is real: when thousands of copiers pile into the same names, they move the prices they're chasing. And eToro's own analysts note the quant methodology is a black box with no user-configurable per-strategy risk caps. Copying the crowd with AI curation is still copying the crowd.

The planner's verdict on the toolbox

Here is how I frame these tools for clients: they are power tools. A power saw in trained hands, used for the right job, with the safety guards on, is genuinely useful. The same saw used as autopilot, or bought because the commercial made it look easy, is how people lose fingers.

The legitimate use case for nearly everything in this chapter is research automation: scanning faster, backtesting ideas honestly, compressing analysis time. The illegitimate expectation, the one the marketing sells, is that the tool replaces judgment, discipline, and a financial plan. It doesn't. Nothing in this chapter replaces those. The next chapter covers the tools that come closest to trying: the robo-advisors and their AI-native successors.