Source: Li, Mai, Shen, Yang & Zhang (2026) — “Dissecting Corporate Culture Using Generative AI”
This is one of those risks that sounds abstract until you imagine explaining it after the fact. Then it suddenly becomes very concrete, very expensive, and very difficult to hide behind a slide deck.
Culture exists in the eye of the beholder — or more precisely, it exists differently depending on whose perspective you’re looking from. This paper, a follow-up to earlier ML-based culture measurement work, uses generative AI to analyse culture across three distinct stakeholder groups simultaneously: managers earnings call transcripts, financial analysts analyst reports, and employees Glassdoor reviews. The core finding: these three groups don’t agree on what a company’s culture is like. And the disagreements are systematic, not random. Different stakeholders assess cultural dimensions differently based on their position, information access, and what they care about. Managers tend to emphasise the dimensions of culture that support their strategic narrative. Employees emphasise the day-to-day experience of working there. Analysts emphasise the aspects most relevant to investment performance. The economic implications of these disagreements are meaningful. When manager-reported culture diverges significantly from employee-assessed culture — when management says “we’re innovative and high-quality” but employees describe a different reality — that gap contains information about future performance. Specifically, the divergence predicts negative operational outcomes. The gap is a signal that management communication is decoupled from organisational reality. This framework helps answer a question that’s been implicit in corporate culture research: whose measurement to trust.
In plain English, that is why the result matters beyond the chart. It changes where people should look, what they should question, and which comfortable assumption probably needs to be retired.
My takeaway: the danger is rarely the dramatic thing in the headline. It is the quiet gap between knowing a risk exists and assigning someone to do something about it. Very unglamorous. Very important.