Masters in Business

Using AI to Find Investing Stories with Perscient Co-Founder Ben Hunt

January 9, 2026

Key Takeaways Copied to clipboard!

  • The core of Ben Hunt's work, both at Perscient and in his academic background, centers on the idea that financial markets are heavily driven by the rise and fall of narratives, which can now be measured with unprecedented scale using AI. 
  • Successful application of modern AI (LLMs) in data analysis requires strict 'context engineering'—constraining the AI to act as a controlled linguistic calculator rather than allowing open-ended, potentially hallucinated responses. 
  • The increasing reliance on narrative construction by central bankers and CEOs, amplified by 24/7 media and smartphones, has made understanding the evolution of these stories a critical, measurable source of investment signal, especially during a 'discovery phase.' 
  • Ben Hunt observes that narratives around US immigration are shifting domestically to be more pro-immigrant, despite current administration policies, which contrasts with international sentiment regarding the end of Pax Americana. 
  • The narrative surrounding private credit exposure is growing in volume and significance, suggesting a potential systemic issue that is not being easily dismissed by financial leaders. 
  • Effective investing and data analysis require understanding the role of chance and probability (the stochastic element) rather than searching for a singular, definitive answer within structured data. 

Segments

Academic Background and Entrepreneurial Bug
Copied to clipboard!
(00:03:38)
  • Key Takeaway: Academia served as a temporary way station for Ben Hunt, who always possessed an entrepreneurial drive stemming from a love for solving games and problems.
  • Summary: Ben Hunt’s academic career, including a PhD from Harvard and tenure as a political science professor, was viewed as a temporary stop. He started his first company while in graduate school and another while a professor, driven by an inherent entrepreneurial bug. He notes that academia’s low stakes often lead to vicious, non-monetary intellectual fights, where the goal is appearing smart rather than being smart.
Academia’s Influence on Investing
Copied to clipboard!
(00:06:44)
  • Key Takeaway: The academic focus on analyzing unstructured data, particularly words and stories, provided a crucial intellectual capital for navigating the investment world beyond structured financial numbers.
  • Summary: Hunt believes his background in academia was beneficial for his later career in investing. He emphasizes that markets are driven not just by structured data (like price and volatility) but increasingly by the words and stories told. His early work involved applying network math techniques to unstructured data, which is the foundation of modern AI and language modeling.
The Rise of Narrative in Markets
Copied to clipboard!
(00:13:22)
  • Key Takeaway: The explicit use of communication policy, or forward guidance, by the Federal Reserve post-2009 crisis catalyzed the overwhelming importance of narrative in financial markets.
  • Summary: The catalyst for focusing on narratives was the aftermath of the Great Financial Crisis, specifically Ben Bernanke’s Federal Reserve using explicit communication to impact markets. Bernanke later admitted that forward guidance became a primary policy toolkit, using words to actively change market behavior, not just communicate policy. This trend has since been adopted by nearly every CEO, as a company’s valuation multiple is essentially a narrative.
Three Secular Changes Amplifying Narratives
Copied to clipboard!
(00:30:18)
  • Key Takeaway: Three structural changes—markets becoming a political utility, the success of narrative construction by leaders, and the ubiquity of smartphones—have made the role of narrative overwhelming.
  • Summary: The power of storytelling in finance has grown due to three secular changes. First, markets function as a political utility, requiring leaders to craft compelling narratives for public support. Second, 24/7 media demands constant opinion and story over hard news, forcing CEOs to master short-form, high-impact communication. Third, personal smartphones create a constant, addictive channel for consuming these narratives.
Perscient’s Data Ingestion and Scale
Copied to clipboard!
(00:33:57)
  • Key Takeaway: Perscient processes over 200 billion tokens (words) from publicly published global sources using established network math, scaled infinitely by modern computing power.
  • Summary: Perscient ingests everything publicly published globally, utilizing data aggregators like Dow Jones and LexisNexis across multiple languages. The firm processes this data using the same core math Hunt used 30 years ago, but at an enormous scale, processing hundreds of billions of tokens recently. This allows them to track the waxing and waning of thousands of specific stories over time, far beyond simple sentiment analysis.
AI as a Constrained Operating System
Copied to clipboard!
(00:43:28)
  • Key Takeaway: Effective use of AI requires constraining the ‘magic genie’ through rigorous context engineering, providing the data and telling the AI how to think to ensure consistent, signal-rich output.
  • Summary: Asking open-ended questions of AI models like ChatGPT is a mistake because they will often hallucinate or provide answers designed to please the user rather than reflect true signal. The key is to ‘stuff the genie in the bottle’ by controlling all inputs and instructing the AI on the exact semantic signatures (the skeleton) it is allowed to analyze. This human-directed process ensures consistency and identifies signal amidst noise.
Descriptive vs. Prescriptive Stories
Copied to clipboard!
(00:47:29)
  • Key Takeaway: Human communication relies on two story types—descriptive (what happened) and prescriptive (how policy should go)—with the latter holding significant predictive capability regarding future actions.
  • Summary: Humans tell descriptive stories, such as reporting a 25 basis point Fed cut, which are factual observations. They also tell prescriptive stories, like stating the Fed should cut rates, which indicate an effort to influence public opinion or policy direction. Identifying prescriptive narratives, particularly in central bank communications, provides predictive capability about intended market nudges.
AI’s Impact on Signal Detection
Copied to clipboard!
(00:55:00)
  • Key Takeaway: Rebuilding Perscient’s software using modern AI as an operating system resulted in an over 100x improvement in signal capture compared to manually constructed small language models.
  • Summary: Early models used by Perscient were like small nets, missing many signals because they could not account for the myriad ways a concept could be expressed. By rebuilding the system using AI for context engineering, they created a much larger net. This technological leap allowed them to capture signals across global languages, such as detecting the collapse in Chinese luxury demand months before it appeared in Western press reports.
Narratives Driving Alpha and Client Use
Copied to clipboard!
(00:59:21)
  • Key Takeaway: Perscient provides clients with data on when the market begins to ‘discover’ a variant perspective or value source, helping investors catch the most profitable ‘discovery phase’ of a trade.
  • Summary: Perscient’s output is not about identifying the absolute truth but about tracking when the broader market begins to recognize a specific story or value source that a client has already identified. This allows clients to time their entry during the lucrative ‘discovery phase’ before a stock or asset becomes widely recognized. The data is also used for financial advisors to proactively address client concerns based on media narratives affecting their portfolios.
Political Narratives and Societal Risk
Copied to clipboard!
(01:06:41)
  • Key Takeaway: The increasing ’narrative violence’ and political polarization measured in domestic media suggest a historical trajectory toward greater domestic political instability and a changing, less attractive meaning for capital markets.
  • Summary: Hunt observes a shift toward ’trench warfare’ in political narratives, which historically correlates with civil unrest and political violence. Furthermore, the data reflects a change in the perceived meaning of capital markets, moving toward speculation and financial nihilism (‘YOLO’). This suggests a troubling future for the role of capitalism if these narrative trends continue unchecked.
Pax Americana and Game Theory
Copied to clipboard!
(01:14:11)
  • Key Takeaway: America First policies risk disrupting the post-WWII equilibrium, potentially reducing economic growth and trade benefits derived from soft power and the dollar system.
  • Summary: The international dimension of the ‘End of Pax Americana’ piece involves game theory where no dominant strategy exists between the US and other nations. Policies like ‘America First’ lead to a different strategic interaction, potentially eroding the economic advantages the US enjoyed post-World War II, such as the reserve currency status. Improvement in the existing system is possible, but the shift represents a fundamentally different set of rules of the road.
Immigration Narrative Shift
Copied to clipboard!
(01:17:02)
  • Key Takeaway: Data shows a significant decline since October in media narratives emphasizing immigration as a problem, replaced by an increase in pro-immigrant sentiment across political affiliations.
  • Summary: Tracking narratives around immigration policy revealed a steady increase in the view that immigration is not working for Americans up until October. Following an event in October, there was a sharp decline in negative immigration discourse, with an enormous increase in pro-immigrant stories, even among Republicans. This shift in public narrative contrasts sharply with the White House policies implemented during the first ten months of the administration.
US Population and Asset Repatriation
Copied to clipboard!
(01:19:35)
  • Key Takeaway: The US population is projected to decrease in 2025 for the first time, coinciding with a clear, slow effort by other countries to repatriate assets away from the US, evidenced by reduced reliance on Treasuries.
  • Summary: 2025 is projected to be the first year the US population decreases due to lower legal immigration and deportations. Data indicates a clear, ongoing effort by central banks and courts outside the US to move assets away from the US, shifting away from Treasuries as a safe haven. While capital flight by US investors is not currently observed, the repatriation of assets is described as a slowly melting iceberg.
Midterm Election Indicators
Copied to clipboard!
(01:22:51)
  • Key Takeaway: Shifts in public narrative regarding the economy and immigration are significant indicators for the upcoming midterm elections, showing an 18% swing in economic sentiment favoring the out-of-power party.
  • Summary: Narratives and opinions are cashed out in elections, unlike markets which cash out daily. The GOP’s standing on the economy has swung from plus 14% two years ago to minus 4% this year, suggesting a substantial shift impacting the midterms. Unpopular administration policies on immigration and tariffs contribute to this shift, though Democratic narratives are not currently powerful or popular.
Ceding Global Narrative Field
Copied to clipboard!
(01:26:46)
  • Key Takeaway: The long-standing American narrative of raising global wealth and lifting people out of poverty is currently non-existent in media, suggesting the US has ceded this narrative field, particularly to China in South Asia and Africa.
  • Summary: The prominent story that American capitalism has raised the global standard of living and lifted people out of poverty is currently absent from public discourse. This narrative has not died but is waiting for a new force to champion it. The US appears to have ceded this narrative ground, especially in South Asia and Africa, to China.
Emerging Credit Story
Copied to clipboard!
(01:29:31)
  • Key Takeaway: The volume of discussion regarding private credit exposure is growing at an unprecedented rate, suggesting a story with significant legs that is unlikely to be contained by simple assurances.
  • Summary: The volume of media discussion concerning alternative asset managers’ exposure to private credit is higher by an order of magnitude than in the last ten years, even surpassing concerns during COVID. This narrative is growing in a way that suggests it will not be fixed by assurances that ’everything is fine,’ similar to past crises where containment proved temporary. This concern about credit has legs and is not showing signs of easing in financial media.
Mentorship and Intellectual Capital
Copied to clipboard!
(01:32:17)
  • Key Takeaway: For those entering academia or early careers in data science or investing, the priority should be building intellectual capital, often best achieved in academic settings before entering roles that require spending it.
  • Summary: Gary King, who pioneered the science of inference, was a major mentor who influenced Ben Hunt’s understanding of data science. Intellectual capital must be built, often in academia, because once one enters roles managing money, the focus shifts to spending that capital rather than accumulating it. This principle is crucial for young professionals in fields like AI or investing.
Hindsight: Process Over Answer
Copied to clipboard!
(01:39:35)
  • Key Takeaway: Looking back, the most helpful realization for investing and data analysis is that success lies in understanding patterns, probabilities, and process, not in finding a singular, definitive answer in the numbers.
  • Summary: The speaker wishes he had understood earlier that there is no ‘Answer with a capital A’ hidden in structured data. The real magic lies in understanding the stochastic element—the role of chance and probabilities. This contrasts with the initial search for a secret formula, emphasizing the importance of adopting a process-oriented approach over seeking a binary outcome.