Creative ways hedge funds are using LLMs

Tracking case studies of investors using LLMs in their research process

Learning from prediction markets

Rick Bhowmick, Head of Data Eng at Coatue, built a system that reads Polymarket and Kalshi and generates a daily investment newsletter based on odds changes

Check out today’s newsletter and here’s the github if you want to try yourself.

My take: I think this is really cool, and modified it to send a daily email to my extended family which has spurred some interesting discussions.

Shareholder activism

Askeladden Capital, a long-only small/micro-cap value investing firm, used LLMs in a proxy battle

During our proxy contest at AstroNova, AI tools helped us produce the extensive ISS deck and other materials that would have been vastly more costly to prepare otherwise (i.e., we would have needed to spend an incremental six figures, and/or developed far lower quality work-product). We earned endorsements from both ISS and Glass Lewis, and the incumbent board requested the CEO’s resignation. I believe, though it is of course impossible to verify, that we may well have run the first AI-powered proxy contest in history.

My take: LLMs should enable more proxy contests against smaller companies that previously weren’t ‘worth it.’ One challenge is the eval on something like this, given these are relatively infrequent and the feedback loop is so long. Will most likely be leveraged by advisory firms and investors with direct experience creating these materials.

Expert interviews

CEO of primary research firm, Kane & Company, uses LLMs to drive more meaningful diligence conversations

Before talking to experts, we run AI queries on the market and competitive landscape. This gives us a baseline of what public sources and models already know.

The baseline serves two purposes. It helps us write better questions because we know the common answers and can push past them.

It also becomes our quality control filter. When an expert’s answer matches ChatGPT’s output too closely, we flag it.

Say for example we asked ChatGPT about the Canadian IT outsourcing market before starting expert calls. It tells us growth was 18 percent. An expert later gave us the exact same number with the same framing. We know to ask where that figure came from and what assumptions drove it.

My take: one challenge with expert networks are “professional experts” - folks who make their living doing calls and are too far away from actual industry to offer true insights. It’s helpful to get a few reps in first with ChatGPT, which is often regurgitating the same content these folks read anyway. Then you’ll be better positioned to push past.

Expert interviews (again):

An AI marketing firm built an LLM system that interviews experts automatically

The only way to capture true expertise is to build an AI interviewer that earns the trust to be seen as a peer. An equal. One that asks questions so insightful the expert reveals the distinctive methodologies they’d normally only share with another seasoned professional. That is the real technical hurdle. Here is how we cleared it. …

The Note-Taker is an internal tool, invisible to the expert, that continuously analyzes the conversation. The Interviewer queries it for structured progress reports.

What the Note-Taker Tracks:

Coverage Analysis: Topics explored with confidence levels (high/medium/low).

Gap Identification: Required areas not yet addressed, prioritized by importance.

Time Status: Pacing assessment against the target duration, with wrap-up triggers.

Pattern Detection: Emerging themes, contradictions, or when the expert defaults to generic “best practices.”

Next Action: A specific suggestion for the next probe.

The key insight is that the Note-Taker returns structured data, not prose. This prevents the Interviewer from getting confused by a second voice. If the Note-Taker flags a gap in “decision-making frameworks,” the Interviewer integrates the suggestion naturally: “You mentioned evaluating channels—walk me through a recent decision where you chose not to invest somewhere.”

My take: I think we’re still a ways away from LLMs running effective investor style expert network interviews. The example in this post is for executives who want to be interviewed, and I think the best investment insights come from human run interviews, ideally unpaid ones with trusted relationships. However, the AI notetaker seems a very useful tool for human interviewers. Could enable folks who are very good at lining up phone calls with experts to conduct them, instead of having to be conducted by expert analysts.

Follow for more case studies of LLMs+Investment Research

If you have any interesting examples or would like to discuss incorporating LLMs into your research process, reply to this email or reach out via X or LinkedIn.