# Why AI Won't Replace Stock Pickers - Jack Woodruff, Candlestick & Vinoo Ganesh, Kepler

## Executive summary

The discussion explores how Artificial Intelligence (AI) will transform fundamental investing, arguing that while AI will automate preparatory tasks and raise the 'table stakes' of data processing, it will not replace the need for human judgment, proprietary research, or the ability to find divergent views. The consensus is that investment management is fundamentally a time management exercise, shifting the focus from data preparation (which AI can handle) to identifying novel, non-obvious information sources. A critical requirement emphasized is the absolute necessity of auditing every number produced by AI models.

## Key takeaways

- AI's Role is in Synthesis, Not Alpha Creation: Long-term stock picking relies on finding a divergent view, which is a human function. AI's primary value is in synthesizing information and automating the preparation work that previously took hours, allowing analysts to focus on proprietary research and channel checks. (2:10, 11:30)
- The Bottleneck Remains Human Judgment: The biggest limitation is that humans are still the bottleneck; information must be processed and internalized by a person. AI helps identify *where* to spend time, but the final judgment and ability to connect disparate data points (the 'investment mosaic') remain human-led. (3:30, 11:30)
- Auditing is Non-Negotiable: It is critical that any model or number produced by AI can be fully audited. The ability to verify the source of data is the most important factor in building trust in automated systems. (7:21)
- Future Roles Shift to Investigators: Future fund teams will require fewer traditional preparers and more individuals with skills akin to investigative journalists or data engineers—people who can conduct deep, proprietary project work. (13:26)

## Technical details

- Data Auditing and Trust: The ability to audit every single number generated by AI is paramount. Using tools without the capacity for full auditability introduces unacceptable risk, even if the automation saves time. (7:21)
- Process Automation and Time Management: AI shifts the process from 'data cleaning and checking' to 'identifying the right questions.' The goal is to use tools to pinpoint areas of high potential for divergent views, rather than simply automating existing workflows like building DCFs or reading 10-Ks. (3:30, 15:14)
- Fund Structure Evolution: Traditional fund structures (pyramid of analysts -> PMs -> CIO) are expected to become more vertical. The focus will shift from high volume of prepared information to high quality, proprietary information generated by specialized roles. (13:26)

## Practical implications

- Focus on building data pipelines that prioritize auditability and provenance over sheer volume of data processing.
- Develop internal tools that function as 'idea generators' or 'research navigators,' guiding analysts to under-researched sectors rather than simply summarizing existing data.
- Re-skill teams away from routine data preparation (e.g., reading transcripts, basic DCF modeling) toward deep, investigative project work and cross-domain synthesis.

## Topics

Artificial Intelligence, Fundamental Investing, Hedge Funds, Data Engineering, Process Automation, Financial Modeling, Kepler, Candlestick Capital Management, AIE CODE summit

Source: https://www.youtube.com/watch?v=y2dyD6slKxE