Need a trading assistant? How about AI?

Open full sized image

Who’s your copilot?

© charles taylor, Yakobchuk Olena/stock.adobe.com; Photo illustration Encyclopædia Britannica, Inc.

With every new type of technology comes both the promise and the reality. Artificial intelligence is no exception to the rule. Not a day goes by when we don’t hear about the amazing things it can do, often followed by a list of its drawbacks and limitations. If you’re a trader, you’ve likely seen no shortage of claims about what AI can do for your performance and profitability and wondered if the hype is real.

The truth is, AI won’t suddenly turn a losing trader into a profitable trader. But what it can do is help you with the grunt work of trading, including some of the repetitive, time-consuming work leading up to a trade. Think of AI tools powered by large language models (LLMs) as an assistant, one who never sleeps, eats, or takes a bathroom break. One who can help you with idea generation, analysis, information organization, and other administrative tasks.

Brokerage platforms have offered stock screeners for years, allowing traders to filter stocks by criteria such as market capitalization, trading volume, sector, and price. AI takes the concept a step further by letting you describe—in plain language—the types of stocks and setups that fit the way you trade. You’re not asking it for “good stocks” to trade, but instead, teaching it to find the best stocks for you, with your style, risk profile, and personality in mind.

Start by giving your LLM a profile. For example, “I am a swing trader looking for stocks that trade at least 2 million shares a day, have above-average volatility, have a history of trending, and are currently in leading sectors.” Or you could say, “I’m only looking for large-cap stocks, but avoid any biotech or pharmaceutical stocks, and make sure the options listed on their shares have enough liquidity for a day trader to easily get in and out of positions.”

Depending on the AI tool and the market data it can access, the criteria are endless (albeit not new). The power of AI is that you can add, subtract, combine, and experiment at lightning speed, allowing you to build increasingly specific screens much faster than was possible before.

You can instruct AI to use technical studies as additional screens. Maybe you want stocks above their 50- and 200-day moving averages, with a relative strength index (RSI) below 60, or names that have recently pulled back to a rising 21-day exponential moving average. You can also have it scan your charts to identify patterns and formations like cup-and-handle patterns, double bottoms or tops, and bear and bull flags.

Keep adding or removing criteria until you’ve narrowed thousands of stocks down to a list that actually fits the way you trade. Once you have your criteria set, you can create a template so future searches can be run with the push of a button, or automated to run on a regular schedule.

A lot of the hype about AI in trading tends to focus on technical analysis. But AI can also act as a research assistant once you have a trade candidate. Here are just some of the things you can put into an LLM:

AI can summarize and evaluate this type of information, but it can also do much more. Instead of asking for an earnings call summary, ask if management’s guidance has changed from the previous quarter and, if so, how. Has their tone changed? What did they say about profit margins and pricing power? Did analysts push back on any of their answers during the Q&A session? Just as with technical factors, the options are endless.

The same process applies when evaluating macro information like Federal Reserve statements, economic data and reports, interest rate data, geopolitical news and events, and government policy and regulatory developments.

Once you’ve narrowed down your trade candidates, you can use AI as a virtual sparring partner to challenge your assumptions and game out what could happen if the variables change.

Give it your thesis and then ask challenging questions. What am I overlooking? What assumptions am I making? What’s the strongest argument against them? What would invalidate this trade? What upcoming events could change the outcome?

AI can also help you compare scenarios. For example, what might happen to my thesis if interest rates rise? Is the trend still valid if it breaks a particular technical level? What if a company’s earnings growth were to slow by 25%?

A word of caution: Don’t make the mistake of thinking that AI knows the answer, or even that there is a correct answer. LLMs have their own biases, and almost everyone who has used AI knows that it can be confidently wrong. Instead, the goal is to force yourself to defend a trade or thesis before you put money behind it. You are looking for weaknesses in your strategy—things that can be obscured by short-term enthusiasm or hidden biases.

Perhaps one of the most powerful but least intuitive ways AI can help with your trading is by identifying patterns, both positive and negative, in your own behavior.

Most brokers will let you download your trade history into a spreadsheet that includes entry and exit dates and times as well as position size, both in shares and dollars. Feed this raw data into an LLM and then query it to surface previously undetected patterns.

What’s your average hold time? Do you do better with swing trades—those held a few days or a week—or are you more profitable when you stick to day trades? What’s the common thread with your winners? Are they more liquid? Are they mostly large-cap? What sectors have you had more success with?

There’s no limit to how granular you can get. Imagine prompting, “Take all of my winning trades and see if there is a pattern as to what time of day I enter and what time I exit them.”

And if you keep a trading journal, AI can take your analysis to another level by reading your entries, either through an upload if they’re digital or a scan if they’re physical. You could then correlate your winning and losing trades with your mindset, emotional state, or levels of stress and anxiety at the time you placed them.

With a little practice and some refinement, you can even use AI to identify the behavioral tendencies that tend to turn a trader into an ex-trader:

Given the right prompt, and a bit of back-and-forth discussion with your AI assistant, you might learn things about yourself that you couldn’t see before.

It would be a cliché to say AI is not a crystal ball. But it’s not a cliché to say that the only limit to how you use it is your imagination. It’s not a substitute for experience, discipline, or judgment. But it can make many of the things you already do as a trader faster and easier, while helping you see things you might otherwise miss.

Think of AI as another tool in your trading toolbox, one that lets you focus on the actual business of trading. How useful it becomes is up to you.