As developers, we’re not just building AI models to analyze companies; we’re stepping into the shoes of active investors. While AI can process vast amounts of data, it misses the human side—the firsthand experiences we gain by touring production facilities or using a product. These insights help us assess a company in ways AI simply can’t.
Today, we’re asking: How important is it to meet a company’s management team when making an investment decision? Are we missing crucial information if we can’t interact with them in person, or can AI fill in the gaps?
The Invisible CEO: Can You Really Know Management from Afar??
Evaluating a company’s management is a cornerstone of investment analysis. Many investors analyze a company’s leadership to assess both the quality of the business and the ability of management to wear two critical hats: being great capital allocators and effective operators. As Warren Buffett famously stated:
“I think you judge management by two yardsticks. One is how well they run the business, and I think you can learn a lot about that by reading about both what they’ve accomplished and what their competitors have accomplished, and seeing how they have allocated capital over time.”
Many investors also examine factors such as incentive structures, equity stakes, and long-term vision to assess leadership quality.
While evaluating management is universally acknowledged as crucial, the necessity of meeting a company’s leadership in person to assess them remains a point of debate within the investment community.
The Meeting Dilemma: A Double-Edged Sword
For some investors, meeting management in person is seen as indispensable for validating integrity, gaining faith in execution, and developing a clear understanding of the company’s strategy. Buffett’s perspective on this is unequivocal:
“We do not find it particularly helpful to talk to managements. Managements frequently want to come to Omaha and talk to me, and they usually have a variety of reasons that they say they want to talk to me, but what they’re really hoping is we get interested in their stock. That never works. You know, managements are not the best reporting parties in most cases. The figures tell us more than a management does. So we do not spend any real amount of time talking to management. When we buy a business, we look at the record to determine what the management’s like, and then we want to size them up, personally, whether they will keep working.”
The Pitfalls of Personal Encounters
While direct meetings with management can provide insight, they come with their own set of challenges that may distort judgment, as Michael Lewis describes in The Undoing Project. Behavioral biases, particularly in the context of decision-making, can skew an investor’s perception. Key pitfalls include:
- Differing Motivations: Investors enter meetings seeking to understand the business and its risks, whereas management views them as sales opportunities. As significant stakeholders, management often emphasizes positives to boost market perception, rarely divulging industry challenges or company risks.
- Confirmation Bias: Many investors unintentionally focus on validating their investment thesis rather than testing it. Management often reinforces this bias by providing optimistic narratives, while investors fail to ask the tougher, risk-focused questions.
- Stockholm Syndrome: Prolonged exposure to management’s charm, intelligence, or perceived integrity can foster an undue sense of trust. This emotional connection may blind investors to red flags, as was famously the case with Elizabeth Holmes and Theranos. Even sophisticated investors overlooked glaring issues due to personal bonds or admiration.
- Self-Deception: Management often believes in the overly optimistic stories they tell investors, distorting facts to protect themselves from harsh realities. This self-deception can mislead investors who rely heavily on management’s guidance.
- No Real Information Edge: Contrary to popular belief, management meetings rarely provide unique insights. Most material information is already public or shared broadly, and any perceived informational advantage may instead serve as a marketing tool for institutional investors.
When Meetings Can Still Add Value
While meetings are fraught with biases, there are scenarios where they can still be valuable such as:
- Filling Knowledge Gaps: These interactions can help clarify business models, unit economics, or industry structures. However, such insights should always be cross-verified with external sources like customers, suppliers, and ex-employees.
- Providing Feedback: Investors can use meetings to influence strategic decisions, such as capital allocation or dividends. For instance, investors have successfully convinced management teams to undertake buybacks, improve dividend payouts, or adapt to industry disruptions.
Our approach is to gather as much diverse information as possible, so while management interactions are useful, they must be approached with caution, always mindful of behavioral biases.
The Rise of the Impersonal Approach: AI and Big Data
While AI can’t replace the human touch—no algorithm can shake hands or sense the nuances of facial expressions—modern statistical tools allow investors to analyze large quantities of data from management interactions, without needing to meet in person. Recorded calls, presentations, and meetings are transcribed into written form for thousands of companies worldwide. When dealing with vast amounts of data, statistical methods excel, as they can efficiently analyze these large datasets to extract actionable insights.
A well-known method for analyzing financial textual data comes from Tim Loughran and Bill McDonald [1]. They created a finance-specific dictionary for sentiment analysis—evaluating the tone and meaning of texts to determine whether they convey positive, negative, or neutral sentiment. Their approach categorizes words into negative, positive, uncertainty, and litigious terms. For instance, words like “growth” and “strong” are positive, while “loss” and “decline” are negative, and terms like “report” and “quarter” are neutral.
While dictionary-based sentiment analysis, like the approach developed by Tim Loughran and Bill McDonald, can be quite effective in financial contexts, it has notable shortcomings. These methods rely on predefined word lists, which can lack nuance and fail to capture context-dependent meanings. For instance, words like risk or liability may have different implications depending on how they are used in a sentence. Risk in “Our company faces significant risk due to market volatility” conveys a negative sentiment, whereas in “We have successfully mitigated risk through diversification,” it signals a positive outcome. However, in a fixed-rule dictionary approach, the word would be treated the same way in both contexts, overlooking these subtle distinctions
To overcome these limitations, modern Natural Language Processing (NLP) techniques leverage allow to analyze financial texts more dynamically. These methods consider word relationships, context, and sentiment shifts within entire sentences or documents rather than relying solely on static word classifications. By using advanced algorithms, NLP can potentially detect subtle sentiment cues, understand linguistic patterns, and provide more accurate insights into financial disclosures, earnings calls, and market sentiment.
Uncovering Hidden Insights: The Role of Earnings Calls
An earnings call is a conference call (typically held in the form of a teleconference or a webcast) during which the management of a public company announces and discusses the financial results of a company for a quarter or a year. Typically these calls consist of a prepared presentation followed by a Q&A session and provide a unique glimpse into management’s behavior, as they reveal more than just financial results—they show how management responds to scrutiny.
- The Prepared Presentation:
This section summarizes the most recent financial period, often supported by polished slides. While informative, these presentations are meticulously scripted, checked by compliance, and designed to emphasize the company’s strengths. Studies show that companies have even tailored disclosures to appeal to machine and AI readers [2]. Consequently, the presentation lacks the spontaneity and transparency necessary to truly gauge management’s character or intentions. - The Q&A Session:
The second part is where the real insights can emerge. Analysts pose questions, and management must respond spontaneously. This unscripted segment offers opportunities to evaluate:- Confidence: How confidently does management handle challenging or unexpected questions?
- Clarity and Specificity: Are answers direct and actionable, or evasive [3]?
- Consistency: Do answers align with previous statements, or are there contradictions?
Though these calls can provide valuable insights, they too have their limitations. Companies can influence the flow of information by for instance prioritizing more optimistic analysts [4], skewing the tone of the conversation.
Also cultural nuances have to be understood as illustrated in our following analysis: The higher the cloud, the more positive the spoken contributions were. American executives tend to be more enthusiastic, while neutrality tends to prevail in Japan (upper plot). The tone of the analysts asking questions is comparable all over the world, which also reflects the internationality, as US analysts also attend calls from European companies and vice versa (lower plot).
By analyzing earnings calls over time, investors can track shifts in management’s confidence or tone, gaining deeper insights into the company’s trajectory.
The Future of Management Evaluation: A Synergy of AI and Human Judgment
The future of management evaluation lies in the fusion of human intuition and advanced AI tools. By analyzing earnings calls and other textual data, AI can offer quantifiable, unbiased insights that would otherwise be subjective. With advancements in audio and video analysis, AI could soon assess vocal cues like pitch changes or stress markers, and even facial expressions, providing a richer understanding of management’s authenticity and confidence.
Imagine a future where AI not only evaluates what management says but how they say it—flagging evasive language, inconsistent narratives, or stress markers in the body language that indicate deeper concerns.
Conclusion: A New Era of Insight
While evaluating a company’s management is crucial for long-term investment success, the necessity of meeting management in person is is discussed controversially. Whatever the personal decision will be: With the rise of AI, investors can now analyze transcripts, earnings calls, and other recorded interactions with greater efficiency, reducing the potential for bias and expanding the scope of evaluation. The use of AI opens up new possibilities for generating investment ideas that would have otherwise been out of reach due to the limitations of attending only a handful of earnings calls in person. AI can process and analyze transcripts from thousands of earnings calls, allowing investors to identify patterns, detect shifts in sentiment, and uncover insights across a vast number of companies. This scalability enables to spot investment opportunities that may have been overlooked otherwise.
As technologies continue to evolve, they won’t replace human judgment but will enhance it—enabling investors to make more informed decisions with a clearer understanding of management’s behavior. In the end, whether through a handshake or an algorithm, the goal remains the same: identifying leaders who will drive long-term value for their companies and shareholders.
4. Cohen, L., et al. (2012). “Playing Favorites: How Firms Prevent the Revelation of Bad News”