Say on Pay in the Age of AI: Separating Signal from Noise

For leaders of compensation planning, the rise of artificial intelligence (AI) creates new unknowns. Proxy voting is one immediate example. This season marks the first year that the full capabilities of generative AI will be available to—and increasingly used by—investors as a supporting tool for their proxy voting process. Use cases may include analyzing proxy statements, preparing engagement questions, or conducting deeper research on custom peer benchmarking, just to name a few.

While AI’s impact in 2026 is unlikely to deliver dramatic shifts or meaningfully move voting outcomes, it will reveal subtle signs of shifts in investor behavior, advisor influence, company disclosure practices, and future norms more generally. For example, will an AI-driven tool or analysis understand the nuances of a one-time equity award that is “special” versus “value-incenting”?

To understand where AI may begin to influence proxy voting, it is important to first consider the baseline. In recent years, Say on Pay (SOP) outcomes have remained relatively stable. From 2022–25, about 70 percent of S&P 500 companies received more than 90 percent support on their Say on Pay proxy advisory vote, according to Farient’s analysis. The number of companies receiving failing votes of less than 50 percent support has slightly declined over the same period, with only 1 percent of S&P 500 companies failing in the past year, and the percentage of companies receiving 50 percent to 90 percent support increased from 28 percent in 2022 to 31 percent in 2025.

Over the same period, the influence of ISS and Glass Lewis has also remained consistent, per Farient. On average:

• An “against” recommendation from Glass Lewis (with a “for” from ISS) results in an 8% decline in SOP outcomes
• An “against” recommendation from ISS (with a “for” from Glass Lewis) results in a 20% decline
• An “against” recommendation from both ISS and Glass Lewis results in a 35% decline

This consistency provides important context for assessing what may, and may not, change this proxy season. That said, several structural factors are beginning to influence how this proxy season is unfolding:

• Regulatory and political pressures have weakened the perceived influence of ISS and Glass Lewis. As a result, both firms have softened the language used in their proxy voting policies. For Say on Pay voting in particular, ISS and Glass Lewis have not changed the fundamental basis of their tests; however, the proxy advisors’ impact may be declining or otherwise diluted as investors take more of this process into their own hands
• Major institutional investors (e.g., JPMorgan, Wells Fargo, and others) are increasingly looking to move proxy voting in-house, explore new AI-empowered external providers, or pursue a hybrid approach. At the same time, investors and their clients are questioning one-size-fits-all voting policies and demanding more customized solutions

However, these shifts are not expected to translate into immediate, large-scale changes in outcomes. From a compensation and SOP perspective, the impact of AI on voting outcomes in 2026 is expected to be limited due to several factors:

• AI tools continue to evolve, and many investors are evaluating whether to build or buy next-generation solutions
• As customization becomes more feasible, investors continue to refine their internal frameworks
• Cultural and governance norms tend to evolve more slowly than advances in technology

That said, incremental changes may surface through observable patterns over the course of this proxy season. Some patterns important to monitor include:

1. Overall voting trends: Will SOP outcomes materially deteriorate?

• Current expectations suggest continued stability, with approximately 90 percent of companies still receiving a passing score (more than 50 percent support)
• However, a greater proportion of companies may fall within the 50 percent to 85 percent range, potentially reflecting increased independent analysis by investors, including the use of recently implemented AI tools

2. Proxy advisor influence: Are ISS and Glass Lewis recommendations still predictive of outcomes?

• Historical patterns indicate SOP results tend to decline by:
  • About 10 percentage points with a negative Glass Lewis recommendation only
  • About 25 percentage points with a negative ISS recommendation only
  • About 40 percentage points when both ISS and Glass Lewis issue negative recommendations
• Negative recommendations are expected to continue influencing outcomes, though their impact may gradually be diluted as investors adopt a mix of internal and external approaches that reflect customized frameworks
• At the same time, some investors may take positions that are more critical than proxy advisors, making it unclear whether outcomes will significantly diverge from historical norms

3. Investor customization: Are investors converging around new norms, or fragmenting further?

Evidence suggests increasing divergence in voting outcomes, particularly in cases where proxy advisors recommend votes “against.”

In practice, customization is beginning to take shape in several ways. Investors are:

• Revisiting historical “checklists” related to year-over-year pay changes, absolute pay levels, and one-time awards
• Translating these checklists into customized policies and actions, influencing voting behavior, engagement preparation, and investor priorities
• Analyzing proxy statements, including extracting key information from the CD&A, reinforcing the importance of how companies frame their disclosures
• Conducting custom peer group benchmarking, enabling comparisons beyond traditional company- or proxy advisor-defined peer groups, with implications for both voting and engagement

Despite these evolving approaches, proxy advisors continue to play a meaningful role in certain areas.

• Investors continue to rely on established pay-for-performance (PFP) alignment models while developing their own frameworks
• Proxy advisors retain advantages in covering non-U.S. markets and smaller companies, where operational complexity remains high

Looking beyond the current proxy season, these dynamics will continue to evolve as AI adoption increases. AI may reshape the proxy voting landscape in several ways:

• Enabling scalable, customized voting frameworks aligned with individual investor priorities
• Reducing reliance on proxy advisors for processing voting volume, particularly in the early stages of adoption
• Making voting behaviors and influence more difficult to predict, but these areas will likely smooth out as investors’ new tools and voting priorities become more clearly defined. This will also largely apply to the top 20 to 30 shareholders for a given company

Farient’s Perspective

In this evolving environment, several core principles remain unchanged for companies navigating proxy voting.

Stick to the fundamentals: Pay decisions should be clearly articulated in proxy disclosure and tied to performance. Investors will continue to view compensation alignment based on proxy table (i.e., grant) values and TSR – companies will still need to monitor and test. Investors are generally going to maintain the same “red flags” or checklists of compensation practices, actions, and P4P misalignments that may trigger a manual review and potential vote against, based on that review and any follow-up engagement.

Prioritize shareholder engagement: As voting outcomes potentially become more fragmented and less predictable, direct engagement becomes increasingly important. Engagement provides an opportunity to add context around complex decisions and industry and company-specific considerations.

Reevaluate peer group strategy: AI tools will enable greater customization in peer group selection, which may diverge from traditional company or proxy advisor frameworks. Companies and their advisors should proactively test a range of peer scenarios and be prepared to address their rationales for the comparisons.

Elevate the Compensation Committee narrative: When relevant and appropriate, the Compensation Committee Chair letter can carry more weight for investors than in the past by providing the full context around the Committee’s decision-making process, in addition to any anecdotes they hear during an engagement. An annual or periodic letter can be a powerful tool. Don’t treat this as boilerplate or a check-the-box exercise; insight from the Compensation Committee Chair and its members can help contextualize non-routine pay decisions, particularly around special or one-time awards. While AI can be helpful for typo and grammar checks, investors have tools at their disposal that allow them to check for AI-generated content. The pen should remain firmly in the hand of the Committee Chair.

---

At Farient Advisors, Angela Moe is a partner and head of consulting services, and Devon Furlong is a director. Ryan Nowicki Stewart is the founder and CEO of Pennant AI, a proxy voting and compliance intelligence platform.

Data for this article was sourced using Farient’s SOP Tracker, which provides five years of proxy voting data on Russell 300 companies that can be sliced and diced by year, sector, and size. Voting results are updated weekly throughout the proxy season by Farient’s Data Analytics Team. Questions about Farient Trackers? Contact bradley.cross@farient.com.