## AI Proxy Advisor for Shareholder Proposals
Overview of This Service
This page provides an AI-driven analysis of shareholder proposals for large companies in the United States (from 2026-02-01). By leveraging custom-built AI workflows, this advisor bridges the gap between academic research and practical, actionable proxy voting insights.
The analysis is powered by a synthesis of two studies:
"Beyond Bias: AI as a Proxy Advisor": Exploring how large language models can provide objective voting recommendations.
"AI Meets ESG: Systematic ESG Classification of Shareholder Proposals": Implementing a framework for identifying and categorizing environmental, social, and governance factors.
From Research to Practice: the Methodology
While the underlying logic is rooted in academic rigor, this service translates the methodology for real-world application:
Foundational studies: utilize simplified, "zero-shot" prompts to ensure reproducibility and to prevent the injection of bias; focus on maximizing accuracy in voting guideline matching through careful manual reviews.
This service: adopts resilient anchors - sophisticated prompting techniques that incorporate contextual data and established governance principles to provide a more dynamic, nuanced analysis; automates end-to-end workflows, ranging from complex text extraction to AI-driven voting guideline matching.
AI Recommendation & ESG Classification
The below tables will be updated periodically as new proxy statements are filed. The second table aggregates the data from the first to illustrate the alignment between actual shareholder votes and the AI’s analysis. While firms may use different vote tabulation bases, I have applied a unified base (totaling For, Against, and Abstain votes) across all proposals to ensure consistency.