Michael Diegelmann: What If Warren Buffett Managed Your Money?
Summary
Closed, rules-based AI scoring beats free-form prompting: fixed inputs yield consistent, comparable company scores instead of hallucination-prone, non-repeatable answers.
Agentic finance — AI research agents scoring corporate filings and feeding automated traders — will make machines the primary readers of company disclosures.
Warren Wise scores companies using Benjamin Graham knockout filters, Warren Buffett quality gates, and Charlie Munger pattern recognition, demonstrated live on Rio Tinto and BHP.
The demo flags gaps between Rio Tinto's narrative and its numbers, while BHP scores as the stronger, cash-rich, consistent reporter.
AI compresses weeks of due diligence into minutes yet cannot detect fraud, so human judgment and management meetings remain essential.
Darrell Thomas's takeaways
- Top takeaway Frames the interview around retail investors doing their own due diligence and warns they must learn AI tools or get left behind.
- Questions whether agentic finance could cut financial advisors out as a fee-charging middleman.
- Presses on coverage, asking whether the tool can analyze juniors on TSX Venture and OTC markets rather than only mega caps like BHP and Rio Tinto.
- Raises the valuation debate, noting SpaceX's initial launch at over a hundred times earnings and asking if multiples still matter.
- Says the dozens of fund managers he has interviewed validate reported numbers in one-on-one CEO meetings and go in well prepared because they do not want to lose money.
Warren Buffett's takeaways
- Top takeaway Presented as the model's core influence: the guest says Buffett's focus on what makes a quality company forms the "Buffett gate" testing the quality and consistency of the quantitative information companies give shareholders.
- The guest paraphrases Buffett's Q&A sessions — more ideas from annual reports than from any sell-side investment-bank research — as the philosophy behind the tool.
- Placed in the Graham-to-Buffett lineage: the guest says Graham's knockout criteria came first, Buffett's quality methodology built on them, and partner Charlie Munger added pattern detection.
- The host points to Buffett's conservative, numbers-first approach as the one that made him one of the richest people in the world.
Michael Diegelmann's takeaways
- Top takeaway Argues agentic finance will fully automate the path from research to trading, with research agents scoring official corporate filings and passing predefined signals to trading agents before markets open.
- Contends credible AI scoring requires a closed, coded system with clear guardrails, warning that open LLM prompting produces hallucinations and results that cannot be compared across companies.
- Explains WarnWise combines Benjamin Graham knockout criteria, Warren Buffett quality checks, and Charlie Munger pattern detection across eight dimensions, scoring companies 1 (knockout) to 5 (premium).
- Asserts machines are now the number-one readers of corporate documents — citing Deutsche Bank's 803-page annual report — and urges CFOs to manage disclosure for machine readers.
- Says an analysis that takes a human analyst two to three weeks takes his system about 35 minutes, while conceding the machine cannot find fraud and final judgment should stay with humans.