JPMorgan Chase terminated its relationships with external proxy advisory firms and replaced them with an internal AI-driven governance system, routing shareholder proposal analysis and voting recommendations through proprietary machine learning infrastructure. The bank holds $3.9 trillion in assets under management across institutional accounts and did not disclose the build cost or deployment timeline, though personnel familiar with the rollout estimate the system went live during the fourth quarter of 2024.
The move affects how JPMorgan analyzes the 400 to 600 shareholder proposals it reviews annually across portfolio holdings. Proxy advisory firms—primarily Institutional Shareholder Services and Glass Lewis—historically provided voting recommendations on governance issues ranging from executive compensation to board composition. JPMorgan now processes ESG disclosures, compensation structures, and board independence metrics internally, applying natural language processing to proxy statements and SEC filings without third-party interpretation layers. The bank confirmed the system cross-references historical voting patterns, regulatory guidance, and sector-specific governance benchmarks, though it declined to specify which large language models underpin the architecture.
This is the first time a top-five U.S. asset manager has fully replaced proxy advisory infrastructure with internal tooling. The decision follows years of criticism that ISS and Glass Lewis wield disproportionate influence over corporate governance outcomes, particularly after studies showed that 25 to 30 percent of institutional investors vote in alignment with ISS recommendations without independent analysis. JPMorgan's shift removes that intermediary and signals confidence that machine learning can parse governance complexity at scale without introducing the conflicts of interest that plague advisory firms—many of which also sell consulting services to the same companies they evaluate.
The timing matters. Proxy season begins in March, and the SEC continues to tighten disclosure requirements around AI use in investment processes. JPMorgan's internal tool sidesteps the regulatory ambiguity surrounding third-party AI vendors while maintaining full audit trails on voting rationale. The bank also avoids the $500,000 to $2 million annual fees that large institutions typically pay for comprehensive proxy advisory coverage. More importantly, it gains speed: internal AI can process amended proxy filings within hours, whereas external advisors often require 48 to 72 hours to update recommendations.
Allocators should track whether other bulge-bracket managers follow JPMorgan's lead. BlackRock and Vanguard remain tied to ISS and Glass Lewis, but both have invested heavily in internal data science teams over the past 18 months. If two more top-ten managers shift to proprietary governance AI by mid-2025, the proxy advisory duopoly faces material revenue compression. Separately, watch for shareholder proposal outcomes in the spring 2025 cycle: if JPMorgan's voting patterns diverge noticeably from ISS recommendations on contentious climate or compensation votes, it will confirm the system's independence and likely accelerate adoption across the allocator class.
The advisory firms issued no comment. JPMorgan's head of stewardship and sustainable finance stated only that the bank "invested in capabilities that enhance our fiduciary obligations." The first test arrives in 42 days when annual meeting proxies begin filing.