David Trainer, CEO of New Constructs, demos FinSights, an AI investing tool built on proprietary financial data and models. He shows it beating the S&P 500 across three Bloomberg-branded indices and contrasts its results with other AI agents like ChatGPT and Claude, which he says give wrong numbers and hide their methodology.
Chapters
Hi, I'm David Trainer, CEO and founder of New Constructs. I'm here to tell you that artificial intelligence for investing has arrived. Introducing FinSights. This is an AI built on Google Cloud by Google Cloud engineers, but based on New Constructs' proven superior data, financial models, and stock market signals. Let's jump right into this and let me show you how it works. Let's start by asking the agent, what are the best stocks in the consumer cyclical sector? Now, we could do this for any sector, but I just wanted to give you one quick example so you can see how quickly and simply this AI delivers real answers. There you go. Those are the best stocks. How about we take a look at a portfolio? Let's say you want to make sure none of your stocks are going to miss earnings in the next quarter. So you ask it, which stocks in my portfolio are most likely to miss? Well, FinSights can give you an answer for that too. And there you go. You can see the stocks with the most earnings distortion. Now let's get even more specific. Why don't we ask it, give me the best stocks in the entire market that also have the best earnings distortion score. And there's the answer. By the way, there's only 10 of them across all 3,300 stocks we cover. That's a pretty good list, folks. Now, how about a pretty good question? Why should you trust these results? What makes New Constructs and Finsights so special? It's about the data, and we've been saying this for years at New Constructs. Models are only as good as their inputs. Garbage in, garbage out. But what goes into the New Constructs models and the Finsights AI is the best data, the best financial models, and the best stock market signals for fundamental data in the world. You don't have to take my word for it. Our research has been vetted by MIT Sloan School of Management, Harvard Business School, Ernst Young, and the Journal of Financial Economics. They've all published papers proving the superiority of our data, financial models, and our stock market signals. Even better than that, we have live traded indices based on our research that are dramatically outperforming the market. We've got the Bloomberg New Constructs Core Earnings Leaders Index, an index based entirely on holding the stocks with the most Core earnings power. The Bloomberg New Constructs 500 Index. That's an enhanced version of the S&P 500. Instead of weighting stocks based on market cap, we weight stocks based on core earnings. And then there's the Bloomberg New Constructs Ratings VA1 Index. The easy name for this? The Very Attractive Stocks Index. Now let's take a look at how these indices perform. First, the Core Earnings Leaders Index. You can see over 5 years outperformed dramatically, up 141% versus the S&P up just 95%. How about just year to date in a shorter time period? Up almost 26% versus the S&P up just 15%. Next, how about we see how the Enhanced S&P 500 Index is doing? And I love this one because this is where we can say, look, we'll let somebody else pick the stocks. We can still outperform because we'll weight them more intelligently. And in this index, we weight stocks based on core earnings instead of market cap. And you can see it drives alpha. It's beating the S&P by over 150 basis points year to date. Over 5 years, beating it by even more. Lastly, let's look at the Very Attractive Stocks Index. You can see over 5 years, this index is dramatically outperforming the S&P 500 as well. The takeaway is the research has real alpha, whether it's the alpha in our core earnings and earnings distortion data, Or the alpha in our ratings. Now let's go back and ask FinSights some more detailed questions. Instead of the big picture questions, let's do a deep dive on a particular company. How about asking for Microsoft's return on invested capital and core earnings over the last 5 years? Nice. There's the data. Let's add some more to that table. There it is. And by the way, if you wanted to download this into Excel, you can do that in just a matter of seconds. Here you can see the data in an Excel file that we just asked for. Let's do another specific question. Let's do a deep dive on A company that's been in the headlines here a bit lately, and look at the difference between its core earnings and GAAP earnings over the last 5 years. Next, let's take a look at how much the agent can explain exactly what it's doing. Let's ask it how to define earnings distortion. There was an explanation with a link into our education system. If you want to know about how New Constructs stock rating system works, as you can see here, The system's completely transparent about exactly how it arrives at its conclusions. We're not here to try to get you to believe in a black box. There's real fundamental insights, data backing everything up that drives these results. Let's get more specific. There are multiple different versions of return on invested capital, and FinSights gives you answers for all of them. FinSights gives you details on how we calculate all of them with links to the source data. So, wondering what else you can get from FinSights? How about credit ratings? Yes, we do have credit ratings, and these are the criteria. How about ratings for ETFs and mutual funds? FinSights can do those too. There we go. In a matter of seconds. Getting the best stocks, ETFs, or mutual funds based on proven superior stock market signals. I think that's pretty strong. Why don't we ask another question? How about the best ROIC in the tech sector? I want to choose an easier question because we're going to compare the answers from Finsights to the answers from Perplexity. Here's what we got when I put this query into Perplexity. It names Apple, AppLovin, Monolithic Power. You see these. Let's see how that Stacks up to what FinCites has: GPT, Nvidia, Apple, Microsoft, Claude, Nvidia, Microsoft, Apple, Mastercard, a few more. We're talking mostly the same stocks from these other agents. What does FinCites show? I don't see anything except for Nvidia on the same list. We're showing much bigger numbers. Why don't we compare the ROIC? That perplexity and Microsoft and Claude have for certain companies to what new constructs has for some of these stocks. So we're not getting a number here for Apple, Mastercard, ROIC, additional high ROIC. We don't really get numbers here from Claude. What does GPT say? GPT gives us some specific numbers so we can say, okay, Nvidia says 176 for Nvidia. We got 180. Not bad. Apple, it says. 30 to 50%. We've got 76%. Not even close. How about Microsoft? 22 to 27. Microsoft? Yeah, 23. Not too far off. Doesn't list any others though. That's too bad. Why don't we see what Perplexity's got? It's got Apple, AppLovin. Let's see, Apple, it says 78. We got it at 76. Not bad. AppLovin, it shows At 52, 62%. Microsoft 28, 23. Texas Instruments, let's add that to the list, 15%. It's got it at 19. And HP doesn't really have a number. I think the point is the numbers from the other agents aren't nearly as good as what you get from New Constructs. They're getting them from all kinds of different places. You don't know if the methodology is the same. You don't know if it's apples to apples. The takeaway is that with Finsights and New Constructs data, you're getting something that you can truly rely on and truly trust. Again, AI is only as good as inputs. If you've got best-in-class inputs driving an AI, you get real insights and real stock picks that you can trust because they have a history of beating the market. Thank you.
