Buffett AI: Building an AI Agent to Think Like Warren | Wharton Online Webinar

Why It Matters

Agentic AI transforms qualitative investment research into a repeatable, button‑press workflow, giving value investors scalable tools while preserving essential human oversight.

Key Takeaways

• AI agents now can execute full investment research workflows with a button
• Agentic era replaces chatbots, enabling deterministic, tool‑calling financial analysis
• Buffett‑style factor models reduce to multi‑factor alpha, questioning durability
• Pattern‑recognition agents can generate qualitative investment ideas at scale
• Wharton program showcases building “Buffett AI” to codify his decision framework

Summary

The Wharton Applied Value Investing webinar introduced "Buffett AI," an experiment to encode Warren Buffett’s investment logic into an autonomous AI agent. Presenter Brett Kin explained how the project fits within the program’s goal of teaching systematic stock‑selection techniques and highlighted the transition from simple chatbots to sophisticated agentic systems.

Kin argued that the current "agentic era" overcomes the stochastic, language‑only nature of earlier large language models. By allowing tool calls to Python, Excel, and APIs, agents can perform deterministic calculations, generate deep research reports, and orchestrate multi‑step workflows through a single button press. This capability is organized via a "skills architecture" that encodes repeatable investment processes.

Concrete examples included an agent producing a 17‑page CEO analysis for Boston Scientific and a diagnostic button that evaluates a company against Buffett‑style criteria such as owner earnings and market‑cap‑to‑GDP ratios. The discussion also referenced academic work deconstructing Buffett’s returns into value, quality, and low‑volatility factors, suggesting that his edge may now be replicated with multi‑factor ETFs.

The broader implication is that investors can now scale qualitative pattern‑recognition—off‑menu to on‑menu opportunities—while still applying human judgment for validation. If adopted widely, such agents could reshape value‑investing curricula and augment portfolio managers, but they do not eliminate the need for rigorous financial modeling.

Original Description

Apply to the next Applied Value Investing Certificate cohort:

https://wallstreetprep.wharton.upenn.edu/applied-value-investing-certificate/?utm_source=youtube&utm_medium=referrals&utm_campaign=fe_avibuffettai_Spring26&utm_content=description_top

Recorded May 1, 2026. The Applied Value Investing Certificate Program webinar series from Wharton Online and Wall Street Prep.

Brett Caughran, Program Director of the Applied Value Investing Certificate Program at Wharton Online and Wall Street Prep, walks through how he built an AI agent designed to think like Warren Buffett, why most AI valuation tools miss what matters most about a business, and where judgment still has to live in the investment process.

About the program:

The Applied Value Investing Certificate is an 8-week program co-developed with Wharton Online. Built for fundamental analysts, portfolio managers, and self-directed investors who want a repeatable process for valuation, quality assessment, and stock selection. Cohorts work through live valuation cases, pitch real stocks, and join office hours with Wharton faculty and practitioners including Howard Marks (Oaktree), Jeff Gramm (Bandera Partners), and Ricky Sandler (Eminence Capital). Graduates close the program in NYC with a cohort photo at the Nasdaq.

Learn More: https://wallstreetprep.wharton.upenn.edu/applied-value-investing-certificate/?utm_source=youtube&utm_medium=referrals&utm_campaign=fe_avibuffettai_Spring26&utm_content=description_bottom

#AppliedValueInvesting #WhartonOnline #ValueInvesting