To create this report, I took the 503 corporate entities listed on the S&P 500 in December 2024.
The quarterly earnings transcripts were then acquired via FinancialModelingPrep's API.
From there, the transcripts were:
- Converted into a structured JSON format
- Processed through Claude Sonnet 3.5 to identify snippets of Artificial Intelligence Discussion
- Categorized into labels to identify if the snippet was describing:
- An company using AI tools for a project
- A company building large scale AI services to sell (Google, Microsoft, etc.)
- A company benefiting from AI infrastructure buildouts (Utilities, Oil & Gas, etc.)
- A company providing AI infrastructure to other companies (Cloud Providers, NVIDIA, etc.)
- Not related to AI
This report largely investigates the category of "Using AI", or companies using AI tools for a project.
After reviewing the initial data, additional prompts were created to analyze and categorize the projects by their User Category, App Types, and Implementation Stage.
With each prompt and review, the data was spot checked for inconsistencies and hallucination. The prompts were always asked to analyze only the data fed via the transcripts and were not allowed to hallucinate or make up data.
ROI
The transcripts were re-analyzed to identify mentions of ROI (Return on Investment) or measured impact from the initiative. Where a mention was found, a score of 1 was applied else null. This score was divided by the count of companies mentioning AI and categorized as a "Using AI" project.
This ROI metric was then used to compare the success of sectors, user categories and others.
While novel, it is limited and relies on what was offered by the companies. Ideally, an expansion of the work performed by MIT Media Labs / Project NANDA could be performed targeting specifically the largest companies and also tracks them over time..
Data Access
The complete dataset used in this analysis is available for download. The data includes 3,017 company-quarter observations across 503 S&P 500 companies from Q1 2024 through Q2 2025.
Dataset Contents
The CSV file contains detailed information about each company's AI initiatives, including:
- Company identification and industry classification
- AI use category and implementation status
- Application types and business units
- ROI metrics where disclosed
- Strategic categorization
File Details: The CSV contains 3,017 observations | 15 columns | 503 companies | 6 quarters (Q1 2024 - Q2 2025)
For Researchers: Please cite this dataset when using in publications. See the data dictionary for methodology details and usage recommendations.