Decoding Corporate AI Readiness: An Open-Data Framework for Scoring Digital Transformation from What Companies Actually Tell Regulators
Abstract
Purpose: Every major corporation now claims to be “AI-powered,” yet no one can reliably measure whether that language reflects real operational capability or just marketing. Proprietary digital-maturity scores are black boxes. Patent counts miss deployers. Job postings capture intent, not execution. This paper cuts through the noise by introducing the AIDI (AI and Digitalization Intensity) framework - a fully transparent scoring system that converts what companies actually tell regulators in their annual filings into a firm-year digital-readiness metric. The index is open-data, replicable by anyone, and designed to work across any industry or filing jurisdiction.
Design/methodology/approach: AIDI works by scanning the business-description and risk-factor sections of SEC Forms 10-K and 20-F for a curated vocabulary of AI and digital-operations terms - spanning core AI methods, analytics and automation applications, and operational digital infrastructure. Hits are normalized per 10,000 words to produce a panel-ready score. The approach is stress-tested through binary, logarithmic, and TF-IDF-weighted alternatives. The framework is demonstrated on ten global oil and gas registrants (79 firm-years, 2015–2024) linked to XBRL-extracted financials, using two-way fixed effects, first-difference models, wild cluster bootstrap, and permutation inference designed for small panels.
Findings: The score reveals a stark digital divide: 63% of firm-years contain zero AI/digital language, while a small cluster of filings show intense disclosure. Among firms that do talk about AI, the ones that talk more have measurably stronger cash-flow efficiency (β = 0.0029, bootstrap p = 0.004) and operating margins (β = 0.0188, p = 0.008). These results survive every robustness test thrown at them - eight alternative specifications, leave-one-firm-out jackknife, and permutation inference all confirm stability. The payoff is especially pronounced during commodity-price turbulence, suggesting that digital capability functions as operational shock insurance.
Originality/value: AIDI is the first open-data, purpose-built scoring framework that turns regulatory filings into a standardized AI-readiness metric. Unlike consultant rankings that cost six figures and cannot be verified, AIDI can be computed by any researcher with a laptop and an internet connection. It works for oil and gas today and is immediately portable to mining, utilities, manufacturing, or any sector with comparable disclosure requirements. The paper provides both the tool and the evidence that it captures something real.