Most Investors Are Using AI Wrong. Here's How To Really Use It.

[00:00:00]Investors using AI today are actually making themselves worse off, and you're probably one of them. You might think AI is giving you an edge, but it could actually be holding you back. Used the wrong way, it could lead to some seriously bad investment decisions. This is something that we're seeing more and more, and what's crazy is how easy these costly mistakes are to avoid. So, stick around because today we're going to show you how most people are using AI wrong and how to use it properly to gain a real edge in your investing. My name is Lewis and welcome to the AI Bureau. The mistake everyone makes goes something like this. They fire up ChatGPT or Claude, type in a stock ticker, and ask something like, "Should I buy this?" or "Where will this stock be in 6 months?"

[00:00:47]They treat AI like a psychic, as if it could comprehend the future better than any human or pick the perfect stock. The problem is, well, it can't. Not sure who needs to hear this, but AI isn't psychic. Large language models predict likely patterns in text based on what they were trained on. They don't have a crystal ball that could see future earnings reports, interest rate decisions, or geopolitical shocks.

[00:01:11]Similarly, stock prices move on information that doesn't exist yet, whether that's a surprise product recall, a central bank decision, or a competitor's product launch. Put simply, nobody can see the future, not even professionals managing billions of dollars in assets. AI is no different.

[00:01:29]Yet, 1 in 10 investors report using LLMs to pick stocks, and that's not even the worst part. Too many investors take AI's answers at face value without ever researching or checking whether they're actually true. But, AI is known to hallucinate. For anyone not already familiar, hallucinations happen when AI generates false information or cites a source that doesn't even exist. Around 20% of AI responses are estimated to contain hallucinations, meaning one in five queries could include incorrect information. And as a not-so-fun fact, 47% of enterprise AI users say they've made at least one key financial decision based on an AI response containing hallucinated data. But even if AI could somehow see the future and pick the perfect stock, it wouldn't stay useful for long. According to a University of Florida finance professor, if AI were genuinely better at stock picking, widespread adoption would simply wipe out its edge. If you want proof that predicting stock prices is genuinely hard, just look at the pros who do it for a living. According to the S&P Global's SPIVA scorecard, around 90% of actively managed US large-cap funds underperformed the S&P 500 over the last 15 years. These are trained analysts with research teams, Bloomberg terminals, and direct access to company management, and yet they can't beat a simple index fund. Why would a chatbot be any better?

[00:02:56]Treat AI like an analyst, not an oracle.

[00:02:59]But if AI can't pick stocks, what is it actually good at? Well, research.

[00:03:04]Instead of asking AI what a stock will do, start asking it to help you understand what a company is, how it makes money, and what could go right or wrong. In other words, treat AI as your own personal research assistant. One of the most powerful things AI could do for investors is read through hundreds of pages of boring documents in seconds. It can summarize earnings calls or explain unfamiliar business models in simple terms. None of that requires psychic powers. It just requires organizing and interpreting information that already exists, which is exactly what AI is built to do.

[00:03:40]This could save you hours of valuable time, which you could then spend on things that are far less tedious. 10-K filings are often hundreds of pages long, and earnings calls can easily run for 45 minutes with 15,000 word transcripts. And let's be honest, even though us dedicated investors find these mind-numbing. Instead of reading through every line, you could ask AI to pull out the details that matter most. You can have it highlight revenue by segment, changes in debt, or new risks the company disclosed. Importantly though, by doing this, you're not skipping the research. But what would have taken you hours can be compressed into just a few minutes. Another thing AI does well is compare companies side by side to see which looks strongest. Two companies in the same industry might seem similar at first glance, but differences often appear when you compare things like margins, growth rates, and debt levels.

[00:04:33]This is something successful investors do regularly, and AI can make those comparisons in a fraction of the time that it would manually take. This is especially useful when comparing three or four companies at once. And if you come across a business model or some jargon that you don't understand, just ask AI to explain it. AI is great at breaking complex concepts down into simple terms, making it one of the fastest ways to go from knowing nothing about an investment opportunity to understanding it well, all in a matter of seconds. AI could help you avoid one of investing's most dangerous traps, confirmation bias. This is when you build an investment thesis and only look for information that supports it while ignoring everything else.

[00:05:15]Usually, that means overlooking red flags and making poor decisions. But you can avoid this by asking AI to actively challenge your thesis. So, AI might not be able to tell you which investments will explode tomorrow, but it can still be incredibly useful. And if you want more tips on how to use this once-in-a-generation tech to its full potential, then sign up for the AI Bureau's newsletter. It's completely free, so what are you waiting for? Just click the link in the description or scan this QR code that's on the screen right now to get signed up. By this point, you understand how both not to use AI and how powerful it could be when used correctly. The question now is, where do you start? Most investors start backwards. They hear a stock or ticker mentioned somewhere and immediately start researching that one company. But this isn't actually the smartest approach because it immediately narrows your scope. A better starting point is to begin with the theme and let the companies follow from there. Ideally, you want a theme backed by real spending, not just online speculation.

[00:06:15]And if you're not sure where to start, well, don't worry. With the right prompts, AI can suggest a range of themes for you to explore. But the important part is using the right prompt. Research has shown that in financial use cases, around 80% of AI output quality is linked to the quality of the input. Put simply, weak, generic prompts produce weak, generic answers.

[00:06:37]The more context you give about yourself as an investor and the type of investments that you're looking for, the better the results will be. Here's an example to show you what I mean. Note that we're using Claude here, but you're free to try this out in whichever LLM that you prefer.

[00:06:52]Anyway, we've given Claude the following prompt. Identify five investment themes likely to see significant growth over the next decade. Focus on structural trends rather than short-term hype. For each theme, explain the key drivers, risks, likely winners, and what would invalidate the thesis. From here, Claude has given us five themes to explore. AI infrastructure, aging demographics and longevity, grid modernization, supply chain reshoring, and water and resource security. Of course, you can ask for more or fewer suggestions depending on what you're looking for or request alternatives if none of these interest you. But since this is an AI channel, let's stick deeper into the first one, AI infrastructure. Thanks to Claude, we could see that the AI infrastructure theme is likely to thrive because of the growing focus on data center capacity, specialized chips, and electricity.

[00:07:45]Claude explains that the winners will likely be companies focused on semiconductors, power equipment and transformers, grid-scale utilities, and cooling and thermal management. We also know the risks. If AI doesn't generate revenue as quickly as expected, it could leave too much infrastructure sitting unused. And if AI chips and models become commodities, profit margins could shrink faster than expected. The AI infrastructure theme could also be invalidated if AI improvements slow dramatically or companies cut spending.

[00:08:17]Once you've picked a theme that interests you, ask AI to highlight companies operating in that space. This gives you a short list of companies to research further. Here, we entered, "Show me 10 companies that fit into the AI infrastructure theme. Include the company's ticker, market cap, annual revenue, and why they stand out." We now have 10 companies, including Nvidia, Taiwan Semiconductor Manufacturing Company, AMD, Eaton Corporation, and several others. Sweet. Once you have a short list of companies within your chosen theme, the next step is understanding each one properly before looking at a single price chart. This is where AI really shines as a research tool. A good place to start is asking what the company actually does. You're not looking for a sales pitch. You just want the basics, like what products or services the company sells. Next, you want to understand how the company makes money. Some companies sell one-time products. Others offer recurring subscriptions, while some earn revenue through fees like advertising or transactions. A company's revenue model determines how predictable the business is, and importantly, how it may perform during a downturn. Ask AI to break down the company's revenue streams and explain which ones matter most. Then, look at who the company's target audience is. A A selling to a handful of large corporate clients carries different risks than one selling to millions of individual customers. Also, check how concentrated its customer base is. For example, if 40% of revenue comes from a single customer, that's probably a red flag. After that, ask AI what factors are driving growth. Is the company adding new customers, selling more to existing ones, raising prices, or expanding into new markets? Each growth engine has a different ceiling and level of risk. And if you're unsure what impacts these factors could have, well, just ask. That's the beauty of AI.

[00:10:10]It can explain things at a pace that suits you.

[00:10:13]The next step is one that most people skip, asking what could go wrong.

[00:10:18]This helps you avoid the confirmation bias that we mentioned earlier. The reality is that every company has risks, whether it's competition, regulation, debt levels, or key employees leaving.

[00:10:29]Ask AI for the company's most significant disclosed risks, pulled from its own filings, rather than speculation. And to really avoid confirmation bias, ask AI to build the strongest bear case against the company that you're considering. Ideally, you should be able to counter it, but if you can't, take that as a sign to pause and rethink your entire investment thesis.

[00:10:50]Doing this for every company on your list can be the difference between understanding what you're investing in and simply gambling on a ticker symbol.

[00:10:57]During your research, you'll find yourself digging through company earnings. But, here's the thing.

[00:11:03]Earnings season produces a flood of headlines, most focusing on one question: Did the company beat or miss expectations? But, this tells you very little. Reading earnings properly means looking beyond the headlines and getting into the nitty-gritty. The good news is that, thanks to AI, you don't have to do this manually. So, start with revenue growth, but remember to look at it in context. A company growing revenue by 20% a year sounds impressive, until you realize that growth has slowed every quarter for the past 2 years.

[00:11:34]That's why you should also ask AI to pull the last several quarters of revenue figures, so you could see the trend rather than just the latest data point. And if you're unsure what the numbers mean, just ask AI to explain everything. Another thing to ask AI to highlight is margins, which matter just as much as growth. A company that grows its revenue while simultaneously losing money on every sale simply isn't sustainable. Ask AI to explain how gross and operating margins have changed over recent quarters, and whether management provided a reason for any shifts.

[00:12:07]Companies often give guidance on what they expect in the near future, and this can matter more to the stock price than the quarter that just ended. For example, a company might beat expectations last quarter but lower its guidance for the next one, causing its stock to fall anyway.

[00:12:22]Ask AI to summarize the guidance from the earnings call and compare it with what management said the previous quarter. And it's also worth paying attention to management's tone, too, not just what they're saying. Confident, specific answers on earnings calls read very differently from vague or defensive ones. Ask AI to highlight any evasive or overly cautious language in the transcript, especially around competition or costs. This could quickly reveal potential red flags. And finally, ask what has changed since last quarter.

[00:12:53]It could be a newly disclosed risk, a shift in strategy, a leadership change, a new competitor, or something else entirely. Companies often bury important updates in dense sections of their filings. Having AI flag what's new can highlight things that you might otherwise miss. All right. So, once you found a company worth investing in, the next step is comparing it with businesses in a similar role.

[00:13:16]A stock could look excellent on its own, but mediocre once you put it alongside its competitors.

[00:13:22]This is one of the biggest blind spots for investors, but thankfully, it's also one of the easiest to fix with AI.

[00:13:29]Naturally, the first thing you'll want to compare is growth. A company growing revenue by 15% a year may sound strong, but if every competitor is growing by 25%, you've probably found the weakest company in that theme.

[00:13:42]Ask AI which companies compete with the one that you're interested in. Then, have it compare their revenue and earnings growth rates side by side.

[00:13:50]Comparing profitability can also reveal which company is being managed more effectively.

[00:13:55]Two companies can generate similar revenue, but if one converts far more of it into profit, you've found the stronger business.

[00:14:03]Margins, return on equity, and free cash flow are all useful metrics to compare side by side, rather than focusing on revenue alone. Valuation is where many investors get tripped up, because a cheaper stock isn't automatically a better one. Ask AI to compare valuation metrics like price to earnings or price to sales, then explain why one company might trade at a premium or discount compared to its peers.

[00:14:27]Sometimes, a premium is justified by faster growth or a stronger competitive position. Sometimes though, it isn't.

[00:14:35]And then, the best way to round out your research is by comparing the risks. One company might carry far more debt, rely on a single major customer, or operate in a more heavily regulated environment than its competitors. A useful tip is to first ask AI to summarize the risks facing the entire theme, like we did earlier, rather than immediately focusing on each individual company.

[00:14:56]Thanks to AI, running this kind of side-by-side comparison takes a fraction of the time that it used to.

[00:15:01]In some cases, it may even change which company emerges as your strongest candidate. So, you've narrowed your list down to a few companies within a particular theme. You know what makes each company tick, you've gone through the financials, you understand the risks, and you're ready to invest. But, there's one final step, one that most people avoid, even though it's what separates good investors from hopeful ones.

[00:15:23]You have your investment thesis, and now is time to challenge it. The first step is to ask what you might be missing. The goal is to have AI identify any blind spots in your reasoning or highlight assumptions that you've made without realizing.

[00:15:38]Sometimes what you've missed is minor and doesn't really matter. Other times though, it can uncover something significant enough to completely derail your investment thesis.

[00:15:47]Of course, every investment thesis is built on assumptions and not everything will play out as expected. So, ask AI where you may be overconfident and which assumptions need to be true for your thesis to work.

[00:15:59]Maybe it's continued to double-digit growth, stable margins, or a key competitor staying weak. Whatever they are, identifying those assumptions makes it much easier to track whether they still hold up over time. Then, finally, take it to the other extreme by asking AI what could make the stock collapse.

[00:16:16]I'm not talking about a sudden 10% drop, but a complete failure of the investment thesis. This could be a failed product launch, a major lawsuit, regulatory changes, or a competitor solving the same problem in a far more cost-efficient way. This step feels uncomfortable because it means deliberately looking for reasons not to invest in something that you were originally excited about. That discomfort is exactly why most people skip it, and that's exactly why you shouldn't. However, if AI can't find a convincing way to challenge your thesis, that's a good sign your investment thesis is solid. At that point, you're either ready to invest or add the company to your watchlist. As you can see, using AI the right way can be an absolute game-changer. It can't tell you which stocks will make you rich, just like it can't predict tonight's winning lottery numbers. Sorry to be the bearer of bad news.

[00:17:05]What it can do is support you throughout your investment journey, helping you make better informed decisions every step of the way.

[00:17:12]So, let's recap this into a simple workflow that you can use. Step one is to ask AI which themes could outperform in the coming years. Ideally, you want to pick a theme that genuinely interests you, whether that's AI infrastructure, energy demand, cybersecurity, defense, robotics, aging populations, or whatever else. Make note of any companies that you think are worth looking closer into.

[00:17:34]Step two is to ask AI to explain what each company actually does, how it makes money, who its customers are, and what drives its growth.

[00:17:44]Don't forget to ask for the bear case as well, helping you avoid confirmation bias early on.

[00:17:49]Once you understand what the business is all about, step three is to check the numbers.

[00:17:54]Pull recent earnings and ask AI to summarize revenue growth, margins, and guidance, and anything that may have changed since last quarter. Step four is to compare your short list against each other directly or against other competitors within the same theme that you're looking at.

[00:18:10]Ask AI to analyze everything side by side, including growth, profitability, valuation, and risk. Step five is to find the bear case before you commit any money. Ask AI for the strongest argument against each company, the assumptions your thesis depends on, and the scenario that could genuinely hurt the investment. Only after going through all of that are you ready for the final step, adding that company to your watch list. This entire process will probably take around an hour, depending on the company, but that's nothing compared to doing it manually. Any experienced investor knows that digging through financial records alone can take hours.

[00:18:48]Add in analyzing the company's fundamentals and comparing it with competitors, and you're easily looking at a full day of research. But with AI, you could do all of that over your morning coffee.

[00:18:59]The important thing though is that at no point are you outsourcing your judgment to a machine. You're simply using it to back up your judgment with better information before you make that winning decision. All right though, that's all for me for today. Now, if you enjoyed this video, show that like button some love and subscribe to the channel for more videos like this in the future.

[00:19:17]Thank you all so much for watching and I'll see you again very soon. This is Lewis signing off.