When senior managers think about how to use AI in the enterprise, discussions tend to go toward cost savings and layoffs. But what if researchers from a major university looked at the financial impacts of AI and, instead, found revenue growth?
Researchers from Carnegie Mellon University (CMU) and Larridin personnel worked together to assess AI adoption more than 500 companies, including nearly all of the S&P 500. (The five largest AI chipmakers were excluded.)
The teams defined seven measures of AI adoption, described below. Then they assessed recent annual filings from the target group.
The results: one measure of AI adoption, called "narrative concreteness," is strongly tied to company revenue growth. This is the first formal study that we know of to find a strong tie between a signal of AI adoption among companies and public company financial results.
"Narrative concreteness" measures how specific the AI disclosures are in the annual reports of Fortune 500 companies. Companies that described more extensive AI adoption and implementation efforts, in detail, with specific results, achieved 8% greater revenue growth than companies that described fewer efforts, and with vague and aspirational, not specific, language.
What was not found was a tie between AI adoption and cost-cutting. More AI adoption was tied to revenue growth across several measures, with narrative concreteness being the strongest. None of the measures of AI adoption were strongly tied to cost reductions, slower growth in costs, nor changes in stock prices.
The CMU and Larridin teams each developed several measures of AI adoption. Then CMU assessed the connection, for each company, between each of the measures of AI adoption and the company's financial results.
Larridin used several sources of publicly available information to develop a total of four scores: AI adoption, AI proficiency, and AI impact per company, plus a combined score labeled the AI maturity index. The CMU researchers developed three additional scores: narrative concreteness around AI and AI investment intensity, derived from required public filings of annual results (known as 10-K reports), and AI-hiring builder rate, derived from publicly available hiring data.
The researchers then compared each of the seven scores per company to three measures of that company's financial performance: revenue growth, operating margins (which reflect both revenue growth and costs), and stock returns. The comparisons were controlled for three factors: industry sector, company size, and prior growth momentum.
All seven AI signals show a relationship to revenue growth. However, the predictive power of most of the signals was reduced significantly after accounting for the three control factors. But narrative concreteness - the extent and depth of information about AI deployments in the 10-K filing - retained strong predictive power even after the control signals were accounted for.
As shown in recent news, AI costs are continuing to rise. Larridin research, and other sources, show that leaders are moving rapidly ahead with AI, while others are standing still. This separation applies to groups of employees, to different teams within companies, and to different companies in each industry sector.
The news also shows that company boards are demanding specifics about where all that AI spending is going. This study provides an overarching data point: during the study period, AI appears to have been growing the top line (revenues), but not yet the bottom line (profits).
How to get ahead of this earthquake of change, even while the ground is still shaking?
Larridin customers manage AI differently than others. They immediately see which AI tools are being used at their companies, licensed and unlicensed, and which teams are using them effectively.
The Larridin platform ties AI usage to workflows and shows our customers where to make improvements. And the platform features specific, detailed features for developers throughout the software development life cycle.
As a result, Larridin customers quickly get the best out of AI: rapid adoption and immediate efficiencies. Best practices from power users are quickly identified, then shared across entire teams. Inefficient spending on tokens is spotted and managed back down immediately, before it becomes a problem.
But, as reflected by the study, many Larridin customers aren't focused on cost reductions. Instead, like others who are benefiting the most from AI, these Larridin customers are thinking about growth.
In this time of rapid change due to AI, companies have an obligation to make the most of it. And to avoid wasteful overspending or missing the boat.
To get access to the full report, click here.
If you're interested in hearing the Larridin story, contact us. We're getting pretty good at telling it.