From UVA Batten to OpenAI: John Bennett on being complementary to change
John “JB” Bennett (B.A. ’14) helps hedge funds, law firms and software companies turn rapidly evolving artificial intelligence into useful tools. The job demands technical fluency but also draws on skills he began developing at the University of Virginia’s Frank Batten School of Leadership and Public Policy: understanding institutions, finding the story in complex information and explaining it clearly to drive decisions.
Now a solutions engineering leader at OpenAI, Bennett works with enterprise customers putting ChatGPT Enterprise and OpenAI’s application programming software to work.
It is technical work, but Bennett does not see generative AI as a fundamental departure from his undergraduate studies in public policy and English literature. “I’m able to do more things with words than before,” he said.
Bennett arrived at UVA knowing he wanted to study English. He had always been, as he puts it, “a book person or a word person.” His future wife encouraged him to apply to UVA Batten’s inaugural B.A. class because of his interest in how government actions affect Americans’ everyday lives.
At UVA Batten, Bennett learned to research unfamiliar subjects, identify how organizations make decisions and turn analysis into a persuasive point of view. Those habits prepared him not only for policy work, he realized, but also for consulting and technology.
“The line between being a policy analyst and being a consultant is not very bold,” Bennett said. Both require understanding a subject in its institutional context, locating the incentives an decision-makers and presenting a compelling case.
His path into technology was less a master plan than a series of choices shaped by circumstance. After graduation, Bennett moved to Seattle when his wife began a master’s degree in public health at the University of Washington. With fewer policy opportunities available there than in Washington, D.C., he entered consulting and began working with technology companies.
The work exposed him to a wide range of problems, but Bennett wanted deeper expertise. A mentor’s career advice that those with complementary skills to a critical resource are always in high demand helped him choose what came next. Storing, analyzing and visualizing data was becoming increasingly central to how businesses operated, so he sought out a role at Looker, a business intelligence platform later acquired by Google. He then spent five years at Google Cloud, ultimately managing solutions engineers focused on data analytics and AI.
The progression from policy to consulting to data infrastructure was not a string of reinventions. Each move added a new capability to the ones Bennett already had.
When ChatGPT was released, Bennett and his brother, a software engineer, began experimenting. They explored a startup concept that would use large language models to support social science research by simulating populations and predicting responses to polls, surveys and market research.
The company never formally launched, but the project changed Bennett’s trajectory. He demonstrated it to former colleagues, including one who had joined OpenAI. The response was immediate: this was exactly the kind of work OpenAI’s customers wanted to do.
Bennett joined OpenAI in August 2025. After years of working on Google’s massive software portfolio, he found the company’s singular focus and relentless pace energizing.
“OpenAI releases something every other week,” he said. “I’ve never seen another company that works like that.”
That speed is shifting the conversation about AI, Bennett said. The question is moving beyond how models perform on abstract measures of intelligence toward what they can accomplish in practice: What can the model do for a person, a team or an organization?
Bennett sees AI primarily as a force multiplier: a way to help people accomplish far more, not to make them unnecessary.
“We’re not trying to replace humans,” he said. “We’re trying to give them six arms” – a metaphor for dramatically expanding what people can research, analyze and create while keeping human judgment and direction at the center.
That view brings him back to the value of complementary skills. Technology alone is not enough; organizations need people who can understand the tools, connect them to human needs and persuade others to act.
One UVA course remains especially relevant for Bennett: a political theory class taught by Professor Paul Martin. Its focus on power, agenda-setting and shaping debate to focus decision-making still guides how Bennett thinks about institutions and influence. The connection to his work today is direct: implementing AI inside a large organization is not only a technical challenge; it also requires understanding who makes decisions, what motivates them and how change happens.
His advice to students is therefore less about predicting the next dominant technology than about committing to lifelong learning.
“The world will always need technical people who can communicate,” Bennett said. “Learn what AI tools can do and what you uniquely can bring to the table.”
