By David Trainer, CEO & Founder of New Constructs
AI tools are increasingly used to summarize companies, screen stocks, and generate investment ideas, but investors can face problems verifying the data and methodology behind outputs.
AI investing tools fall into two categories: systems that generate fast financial commentary and systems that produce verifiable analysis. New Constructs’ system allows users to trace investment conclusions back to source SEC filings, review key assumptions, and understand how insights are produced.
1. What are the biggest dangers of relying on AI-generated investment recommendations when the underlying data sources and methodologies remain opaque?
The biggest danger is that investors mistake confidence for accuracy. Relying on decisions or trades based on bad or unreliable data is a good recipe for making misinformed investments.
Most AI systems are incredibly good at producing persuasive answers, but if you don’t know where the data came from, how it was processed, or how the conclusion was reached, you’re essentially trusting a black box with your money.
In investing, bad data leads to bad decisions. If an AI recommendation is based on incomplete financial information, flawed assumptions, or unreliable sources, the output may sound convincing while also being fundamentally wrong. That’s a dangerous combination because investors often don’t realize there’s a problem until after they’ve lost money.
The reality is simple: If you can’t see the data and you can’t understand the methodology, you have no way to assess whether the recommendation deserves your trust.
2. How can investors distinguish between AI that is merely generating persuasive answers and AI that is delivering genuinely reliable financial analysis?
The first question investors should ask is: “Show me the data.”
Reliable AI should be able to identify its sources, explain how the data was processed, and provide a clear path from the underlying facts to the final conclusion. If an AI provider isn’t willing to show you that information, you should ask why.
Persuasive AI focuses on producing an answer. Reliable AI focuses on producing a verifiable answer.
At New Constructs, we believe transparency is the ultimate test. We also trust that smart investors will read the footnotes and click-throughs of the sources we provide. We show investors exactly where our conclusions come from because we’re confident in the quality of the work. If you can’t inspect the inputs, you shouldn’t trust the outputs.
3. Why is explainability becoming one of the most critical requirements for AI in investing, and what risks emerge when it is absent?
Explainability is becoming critical because investors are learning that AI isn’t automatically reliable.
The more we use AI, the more we understand that the more important the decision, the more important it becomes to understand how the recommendation was generated. Investment professionals, fiduciaries, and institutional investors have a responsibility to conduct proper due diligence. They can’t do that if they’re relying on a system they don’t understand. It’s like in 6th-grade math, when the teacher asked you to show your work – we need that proof.
Without explainability, investors are exposed to hidden biases, flawed assumptions, and data quality issues they can’t detect until it’s too late.
In my view, recommending a completely opaque AI system for investment decisions is no different than recommending a black-box investment strategy without reviewing the underlying holdings. Responsible investing requires transparency.
4. How significant is the threat of AI hallucinations in financial markets, where even minor errors can have major consequences?
The threat is very significant because financial markets are uniquely vulnerable to misinformation.
A small hallucination isn’t always a small problem. An incorrect earnings figure, a fabricated source, or a misunderstood accounting adjustment can dramatically change an investment thesis. Once that misinformation spreads through automated systems, it can influence thousands or even millions of decisions.
What concerns me most is that many AI systems are optimized to provide answers rather than admit uncertainty. In other words, they may generate something plausible instead of acknowledging that they don’t know.
That’s why trustworthy financial AI must be grounded in verified data and transparent analytics. In investing, being confidently wrong can be far more dangerous than admitting you don’t have an answer.
5. How does New Constructs’ Robo-Analyst technology ensure that investment conclusions are grounded in financial facts rather than assumptions, predictions, or potentially unreliable external content?
Our Robo-Analyst starts with a fundamentally different philosophy: facts first.
Instead of relying on unstructured, unverified internet content or broad web searches, our technology analyzes company filings, footnotes, and management disclosures to build a structured financial knowledge base. Every data point is organized within a rigorous taxonomy and ontology designed specifically to measure economic earnings, profitability, valuation, and investment risk to more accurately predict stock prices.
The result is that our conclusions are anchored in audited financial disclosures rather than assumptions or opinions.
What’s especially important is that the quality of our data and models has been independently validated by leading institutions, including Harvard Business School, MIT Sloan, Ernst & Young, and the Journal of Financial Economics. Our research also powers Bloomberg indices that have outperformed the broader market.
Better analysis starts with better data, and that’s where our competitive advantage begins.
6. How does FinSights AI demonstrate the importance of combining advanced AI capabilities with high-quality, structured financial data?
FinSights is a perfect example of why data quality matters more than anything.
Google selected New Constructs as the exclusive data provider for FinSights because they wanted to demonstrate the power of their AI when endowed with highly predictive, independently validated financial data.
The lesson is straightforward: AI is only as good as the information it’s given. You can have the most sophisticated AI model in the world, but if the underlying data is incomplete, inaccurate, or poorly organized, the answers won’t be reliable.
Our AI input is based on more than 20 years of financial filings. FinSights shows what’s possible when you combine state-of-the-art AI capabilities with an extensive and detailed financial dataset built specifically to identify economic reality, not just reported accounting results. That’s how investors get information they can actually trust—not because the AI is smarter, but because the foundation underneath it is stronger.
About David Trainer:
David Trainer is a Wall Street veteran and corporate finance and artificial intelligence expert. As CEO and founder of New Constructs, he specializes in building AI-driven technologies that collect data directly from financial filings, including the footnotes, and building proven-superior stock ratings, financial analytics, and valuation models. He is the author of Modern Tools for Valuation (Wiley Finance, 2010). His company, New Constructs, leverages proprietary AI-driven Robo-Analyst technology to deliver proven-superior stock ratings, investment research, earnings analysis, cash flow and valuation models on over 10,000 stocks, ETFs and funds.
About New Constructs:
New Constructs is an independent financial technology firm bringing transparency and accountability to investment research. Combining accounting expertise with patented AI technology, the company analyzes SEC filings and financial disclosures to uncover the true profitability and valuation of public companies. Its Robo-Analyst platform automates financial modeling and investment ratings across more than 10,000 securities, while FinSights, built with Google Cloud technology, demonstrates how AI can support investment analysis when grounded in trusted, auditable data. For more information, visit https://www.newconstructs.com/
