The Edge Has Shifted | Matt Reustle on How the Best Investors Use AI

Key Concepts

• AI Evolution: The shift from basic LLMs to powerful “agentic AI” is transforming investment research and workflows.
• Leverage & Productivity: AI’s primary benefit lies in augmenting human capabilities, increasing productivity, and enabling broader information coverage.
• Prompt Engineering: Effective AI utilization requires skillful prompt development and iterative refinement.
• Build vs. Buy: Leveraging existing AI tools is generally more efficient than in-house development, especially for firms lacking substantial engineering resources.
• Generalist Empowerment: AI levels the playing field, enhancing the capabilities of generalist investors.
• Customization is Key: The future of AI in finance lies in highly personalized tools tailored to individual needs.

The Rise of Agentic AI in Investment

The conversation traces the evolution of Artificial Intelligence (AI) in the investment world, moving beyond the initial hype surrounding Large Language Models (LLMs) like ChatGPT – initially viewed as advanced search engines – to the emergence of “deep research models” in late 2024/early 2025, coinciding with premium subscription tiers. A crucial distinction is made between LLMs, described as “calculators” performing single-shot tasks, and “agentic AI,” which demonstrates reasoning, contextual understanding, and the ability to synthesize information from multiple sources. This agentic AI is considered significantly more valuable for professional investment work.

Practical Applications & Workflow Integration

AI offers substantial leverage in due diligence and research. For example, tasks like tracking customer commentary on logistics costs during earnings season, previously requiring manual transcript review, can now be automated. AI can also rapidly generate summaries and analyses, effectively replacing the delays associated with traditional sell-side research. Companies like Portrait Analytics are highlighted as leaders in applying AI to investment analysis, consistently identifying emerging trends. Beyond research, AI can monitor sectors, companies, or KPIs, providing nuanced alerts beyond simple keyword notifications (like those offered by Bloomberg). Successful AI users adopt a systematic approach: testing models with detailed prompts, comparing outputs, and refining prompts based on results.

Mastering the Art of Prompting

Developing effective prompts is central to AI utilization. The process involves:

1. Defining the AI’s “role” (e.g., “junior analyst”).
2. Specifying the target audience (e.g., investment committee).
3. Clearly stating the request’s purpose.
4. Indicating preferred information sources.
5. Defining the desired output format. Iterative refinement, asking the AI to identify areas for improvement, is crucial.

The “Build vs. Buy” Debate & Cost Considerations

A recurring theme is the “build vs. buy” dilemma. The speakers caution against the frequent failure of firms attempting to build AI tools in-house, emphasizing the significant resources – both time and financial – required for successful development. Even basic applications can become complex and expensive. The cost is measured in both time investment and “tokens” (the computational units used by LLMs). Leveraging off-the-shelf solutions is generally recommended unless a clear, justifiable need for custom development exists.

AI, Alpha, and the Changing Role of Investors

While a direct correlation between AI usage and alpha generation isn’t immediately apparent, the initial value lies in increased productivity and efficiency. This shift is comparable to the adoption of tools like Bloomberg terminals, calculators, or Excel – essential components of modern investment practice. AI’s ability to rapidly process and synthesize information empowers generalist investors, reducing the barriers previously faced and potentially closing the gap with specialists. However, specialist knowledge remains valuable, though its relative advantage may diminish. The speakers acknowledge the importance of humility, recognizing that even with AI, analytical processes can lead to unintended conclusions.

Future Trends & the Rise of Customization

Looking ahead five years, the speakers predict a significant trend towards customization of AI tools. Current AI outputs are largely standardized, but the future will likely involve highly individualized models tailored to specific user needs and preferences. “Agentic AI” – AI systems capable of performing tasks autonomously – is identified as a key area of development, envisioned as a “human-like assistant” providing rapid access to information and insights. The speakers also highlight the potential of “coding agents” to lower the barrier to entry for software development, enabling users to create tools previously inaccessible.

Adoption & Existential Concerns

While an immediate existential threat to investment professionals is considered overstated, failing to utilize AI will likely lead to reduced productivity. A cultural divide exists in adoption, with the tech industry embracing AI while some traditional financial organizations remain skeptical. Early adopters are expected to gain an efficiency advantage.

Conclusion

AI is rapidly evolving from a promising technology to a practical tool for investment professionals. While not a replacement for human judgment, it offers significant leverage in research, analysis, and workflow efficiency. Successful integration requires a systematic approach to experimentation, a focus on prompt engineering, and a pragmatic assessment of the “build vs. buy” dilemma. The future of AI in finance lies in customization and the development of agentic systems that augment human capabilities, ultimately empowering investors to make more informed decisions.