AI Stock Picker Apps: What They Actually Do
Apps that promise AI stock picks fall into three groups, and they are priced as though they were the same product. They are not.
The three groups
| Type | What it really does | Typical price |
| Screeners with a ranking model | Sorts listed stocks by factors — momentum, valuation, earnings revisions — and presents the top of the list | Monthly subscription |
| Research summarisers | Condenses filings, earnings calls and news into briefs | Monthly subscription, sometimes free tier |
| Managed portfolios | Actually invests, usually as a registered adviser, using a rules-based allocation | Percentage of assets |
Only the third manages money. The first two sell information and leave every decision, and all the risk, with you.
On ranking models
A factor screener is a legitimate tool. What it is not is a prediction: sorting by momentum tells you what has already risen. Backtested performance shown in marketing was produced by running the rules over history that already happened, with the rules chosen because they worked on that history.
If an app publishes performance, look for whether it is live and audited or backtested, and whether it includes the picks that were removed.
Five questions before subscribing
- Is the firm registered? Anything giving personalised advice should be a registered investment adviser. Check it.
- Is performance live or backtested? The gap between the two is usually the entire pitch.
- How is the app paid? Subscription, payment for order flow, affiliate commission on broker signups — each creates a different incentive.
- What happens to a pick that fails? Track records that quietly drop losers are not track records.
- Would you follow it? A screener suggesting twenty trades a month is a tax and cost problem before it is an investing strategy.
When one is worth paying for
A research summariser earns its fee if you read filings and want the reading time back. A managed portfolio earns its fee if it stops you from making changes you would regret. A pick-generating screener has to beat an index fund by more than its subscription cost plus the tax on the extra trading — a harder test than it sounds.
Related
The safer end: robo-advisors compared. The dangerous end: AI trading bot warning signs.
Doing it yourself: ChatGPT for investment research.
Common questions
Do AI stock picker apps actually beat the market?
Most publish backtested rather than live, audited results. To be worth it, picks must beat an index fund by more than the subscription plus tax on the extra trading.
Are AI stock picking apps regulated?
It depends on what they do. Apps giving personalised advice or managing money should be registered investment advisers; screeners and news summarisers often are not.
What is the difference between an AI stock picker and a robo-advisor?
A robo-advisor invests and rebalances for you at a percentage fee. A picker sells you information and leaves the decisions and the risk with you.
Is backtested performance meaningful?
Rarely. It is produced by applying rules to history already known, with rules chosen because they worked on that history.
Figures and rules in this article are checked against the official sources below. Where they disagree with us, they win.