How Investors are using AI - [Business Breakdowns, EP.240]

An effective mental model for prompt engineering is to imagine writing an email to a smart junior analyst working overnight. You must provide the task, the context behind it, desired output format, and specific guidelines, assuming they have intelligence but no background on your thinking.

Instead of just tracking hard numbers, AI tools can systematically analyze years of transcripts to map out qualitative or "soft" guidance (e.g., "revenue will accelerate in H2"). This creates a picture of a management team's guidance style and credibility, a crucial but historically painstaking analysis to perform.

The "memory" feature in today's LLMs is a convenience that saves users from re-pasting context. It is far from human memory, which abstracts concepts and builds pattern recognition. The true unlock will be when AI develops intuitive judgment from past "experiences" and data, a much longer-term challenge.

Firms that meticulously document the reasoning behind trading decisions are building a proprietary dataset for future AI agents. This intellectual property, capturing the firm's unique philosophy, will be invaluable for training AI that can truly understand and operate within its specific context, forming a powerful competitive advantage.

To stay on the cutting edge, maintain a list of complex tasks that current AI models can't perform well. Whenever a new model is released, run it against this suite. This practice provides an intuitive feel for the model's leap in capability and helps you identify when a previously impossible workflow becomes feasible.

Beyond simple quantitative screens, AI can now identify companies fitting complex, qualitative theses. For example, it can find "high-performing businesses with temporary, non-structural hiccups." This requires synthesizing business model quality, recent performance issues, and the nature of those issues—a task previously reliant on serendipity.

Agentic AI is most advanced in software engineering because code provides a constrained, text-based, and verifiable environment. AI agents can now operate for hours, understanding codebases and fixing errors. This iterative reasoning process is a direct preview of how AI will eventually perform long-running, complex investment research tasks.

AI's primary value in pre-buy research isn't just accelerating diligence on promising ideas. It's about rapidly surfacing deal-breakers—like misaligned management incentives or existential risks—allowing analysts to discard flawed theses much earlier in the process and focus their deep research time more effectively.

Forcing investment professionals to adopt specific AI tools often backfires. An investor's research process is deeply personal and tied to how they build conviction. Successful adoption happens bottoms-up, where individuals find tools that reduce friction without compromising their unique workflow for developing trust in an investment thesis.