Me and My Agents — SkimNews
Get the Finance newsletter
Daily finance — markets, central banks, M&A, the prints that move money. Free.
- Four AI agents named Ernie, Thelma, a Mike Mayo bot, and an Excel modeller were trained by the author to parse earnings releases, track investment themes, preview analyst questions, and update financial models for the six largest US banks' Q3 reports.
- Q3 bank earnings produced a combined 250+ pages of filings and 8+ hours of conference calls, with JPMorgan alone furnishing 523 new data points for investors to absorb.
- The Mike Mayo bot previews the questions Wells Fargo Securities analyst Mike Mayo—who has covered banks through 140+ earnings seasons—would ask on earnings calls.
- The agents' read was more bullish than the market's: guidance was broadly up, particularly on net interest income, while equities trading and banking fees offset weaker fixed income trading.
- The author notes earnings release dates account for just 1.6% of trading days but drive roughly 15% of annual share-price variance, making the speed advantage of AI parsing valuable.
- Paid subscribers can email the author for access to the instructions written for each agent and their actual outputs from this quarter's bank reports.
Why it matters: Six of the largest US banks dumped over 250 pages of filings and hosted 8+ hours of calls in one week—and the author's AI agents processed them while he brewed an espresso. If a single investor running custom bots can outpace the consensus read on NII guidance and trading revenue splits, it signals a meaningful speed advantage in an event where 1.6% of trading days drive 15% of annual share-price variance.
Ask SkimNews
