AI Is the Best Friend of Shareholder Activists

Why It Matters

AI‑driven voting tools could democratize proxy decisions and disrupt the proxy‑advisor market, altering outcomes of activist campaigns and corporate governance battles.

Key Takeaways

• J.P. Morgan replaced proxy advisors with proprietary AI for voting decisions
• AI recommended activist support 37% vs >50% from Glass Lewis/ISS for insurers
• AI relies heavily on press releases and volume‑driven digital content
• Biases differ from traditional advisors, leading to divergent vote outcomes
• Firms must redesign narratives for algorithmic interpretation, not just human persuasion

Pulse Analysis

The rise of large‑language‑model AI in shareholder activism marks a watershed moment for corporate governance. Historically, institutional investors have leaned on proxy‑advisor firms such as Glass Lewis and ISS to parse dense filings and recommend votes. J.P. Morgan’s decision to replace those services with a proprietary AI engine signals confidence that machine‑learning can synthesize vast data sets faster and more consistently. This shift not only reduces reliance on costly advisory subscriptions but also introduces a new layer of algorithmic judgment that can sway voting outcomes across the market.

KekstCNC’s analysis of contested votes between 2023 and 2025 reveals that AI‑generated recommendations diverge sharply from traditional advisors. While AI showed a 37% support rate for activist cases involving insurers, Glass Lewis and ISS recommended support in more than half of those instances. The discrepancy stems from AI’s heavy weighting of owned content—press releases, earnings calls, and other high‑volume digital signals—over nuanced qualitative assessments. Consequently, AI exhibits distinct biases, often favoring activist proposals that generate louder online discourse. This dynamic could empower retail investors who access AI tools, but it also raises concerns about the quality of information driving institutional votes.

For corporations, the implication is clear: communication strategies must evolve beyond conventional media outreach. Narratives need to be crafted with algorithmic consumption in mind, ensuring that key messages are both human‑persuasive and machine‑readable. Companies should prioritize concise, data‑rich releases and consider SEO‑friendly phrasing to influence AI aggregators. Meanwhile, regulators may need to examine the transparency of AI voting models to safeguard against systemic bias. As AI continues to embed itself in proxy voting workflows, firms that adapt their disclosure practices will be better positioned to shape favorable outcomes in an increasingly digital shareholder landscape.