Stanford GSB’s If/Then podcast features an interview with one faculty member each episode. Hosted by Kevin Cool, the podcast dives into the findings of researchers and considers how they might apply to today’s big business-world questions.
We’re taking a look at one particular episode, titled “The Data-AI Train Is Leaving the Station,” in which Cool interviews Amir Goldberg, a Professor of Organizational Behavior from Stanford GSB. Why? Because Professor Goldberg’s findings have astonishing implications for the world of business and management.
The Research
In this episode, released on May 28th, 2025, Cool introduces Goldberg as a professor who studies culture in a “novel way:” or, as Cool describes it, with a “data driven approach.” Professor Goldberg goes on to explain his research by saying that there are “many, many, many… processes” that make up culture, and that his work distills them into measurable data.
One area of Professor Goldberg’s research that is particularly relevant to business and management is innovation. Innovation is tricky, he says, for a number of reasons.
First, because we want to know how to increase innovation. And to know that, we need to know where innovation is coming from. There is a longstanding debate, he describes, over whether innovation stems from the margins or from the center of society. The innovation-at-the-fringes theory paints the inventor as a disruptor with nothing to lose, harnessing their outsider perspective and aligning with American individualist ideas. Innovation-at-the-center, conversely, considers the power and resources that those inside society have to innovate and to influence the perceptions of others.
Second, and building on the first, because locating innovation becomes harder when it is not obviously measurable. Groundbreaking developments that occur in technology are easy to track; those that crop up in areas such as strategy or creativity are less so. Professor Goldberg gives an example here: Walmart, which changed retail as we know it by scrapping features that existed “by virtue of convention,” “reimagining what is desirable” and “thinking outside the box.” These changes were strategic, un-patented and hard to measure—but no less impactful for being so.
So what does Amir Goldberg do in order to measure and locate the source of immeasurable innovation?
Transforming Cultural Information into Data
Part of the answer comes down to looking at how people speak. In a “really exciting moment of computational innovation,” Professor Goldberg’s research takes “how people speak and transform it into data.”
By “people,” he means politicians and executives; his research analyzes congressional records, descriptions of strategy, quarterly earning calls, the full judicial record, and more. There are, as is clear from this glimpse of features to be analyzed, “many, many, many millions” of useful examples, all of which are “immensely unstructured.”
This is where AI comes in. AI can absorb huge quantities of data, identifying patterns in a way that humans never could. (Interestingly, Professor Goldberg explains, the human brain does sort of do this—but the process is largely intuitive and not numerically translatable).
Using this technology, Professor Goldberg’s team analyzed speech across multiple contexts to discover who was producing innovative ideas. Their conclusion? Innovators exist on the fringes.
They are, Professor Goldberg elaborates in the podcast, “politicians who are not at the center of … Washington…. people who start firms that are small… judges who operate at the lower courts, who are not trained in prominent schools or by prominent justices.”
What Does This Mean for Organizations?
Locating Innovation
The ability to locate the source of innovation is crucial information for effective management. Many organizations ask themselves why they are behind on innovation and creativity compared to their competitors; to answer this question, they first must understand if innovation is happening already in their company—and they are simply missing or suppressing it.
Companies can do this using a similar process to Professor Goldberg’s speech analysis. Perhaps without realizing it, they are already collecting reams of data on experiences within their organization: employee conversations, documents, idea-exchange over Slack.
Companies can transform these experiences into data using contextual embedding models and large language models. Once done, they can analyze the data to discover where innovation is happening and if it is being allowed to develop. Leaders need to ask “Where are the good ideas happening in my organization? Who’s killing them?” and “Why?”
Professor Goldberg can’t offer the solution, but he can promise this: leaders will struggle to come up with a solution if they don’t know exactly what the issue is and where in the organization it is happening. They need to “diagnose” it: and for that, they need data.
Simulating Responses and Reactions
There is another way that businesses can harness AI to improve. Once again this comes back to Professor Goldberg’s speech analysis research—or rather, to host Cool’s follow-up question:
“This is all really interesting, the idea that you can sort of use data to create almost sort of a profile of a cultural context. Can we take it one step further? Can AI tell us about the way humans operate at some level?”
Professor Goldberg explains that AI might be able to do something along these lines: by predicting how humans will respond and react to company policies. He names computer scientist colleague Michael Bernstein: Bernstein, he explains, is using LLMs to transcribe an interview, feeding that interview to a machine, and asking the machine to predict, based on that interview, how the interviewed person would respond to a series of questions. The results are, he says, “astonishing,” and can be “very consistent with what that human would have answered.”
The utility of this for organizations is clear. Simulations could be run on policies to predict the responses of employees, clients, customers and more, allowing information to be gathered about efficacy before implementation.
How Important is This for the Business World?
The potential impact of AI here is huge. The ability to transform experiences into data could give leaders never-before-seen insights into the way in which they are managing their organizations. Using data is “the imperative of leadership today,” notes Professor Goldberg—”the data/AI train is leaving the station,” and those not on board will fall behind.
There are two things that Professor Goldberg takes care to note when discussing the transformative power of these processes.
The first is, of course, the ethical implications. Transforming employee conversations, interactions and actions into data has an impact that must be considered—one so potentially complex that both Professor Goldberg and Cool decide another episode dedicated to tackling the subject would be best.
The second is that data is “not beneficial in and of itself.” Data aids leaders in identifying where the problems lie in their organization; but human leadership and strategy is still needed to fix those problems. It is people who transform the data into change, and into a competitive advantage.
“At the end of the day,” Professor Goldberg concludes, “that is good news.” AI technology won’t replace leaders; it will simply aid them in gathering the data they need to make strategic, innovative decisions.
