I’ve just started a video call with a general partner at a venture capital firm. “Whenever you’re ready, please start your pitch,” he says.
For a reporting experiment, I am playing the role of a hopelessly delusional tech founder. As I outline my plan to build an “Uber for bananas,” his calm, polite smile doesn’t move. Neither do the small shadows cast by the drape of his shirt. His shoulders hover back and forth just slightly – and a little too consistently. Because while it looks like I’m talking with Jeff Bussgang, the co-founder of Flybridge Capital and a senior lecturer at Harvard Business School, I’m actually talking with his AI-generated video avatar.
When I speak with the real Bussgang – who, much like his AI counterpart, is not impressed with my idea for faster banana delivery – he concedes that the simulation is at least a little “creepy.” But, he says, “My students love it.”
Bussgang is one of several teachers at Harvard Business School who volunteered to help create AI clones of themselves for a new eight-week entrepreneurship boot camp called HBS Foundry. Harvard has tested the product, which costs US$699 (RM2,814), with more than 100 universities across 50 states.
Foundry guides participants through a framework for starting a company, from refining the idea to pitching it to investors and customers. Along with Harvard Business School content, its virtual work space includes a community of peers and access to AI agents – some intended to “represent” instructors. It also features video call simulations for practising pitches, sales calls and board meetings with instructor AI avatars, like the one of Bussgang.
Experienced founders, investors, and Harvard Business School faculty lead weekly live sessions, but the purpose of the online portal is to guide students through AI resources and content from the business school for building a startup.
The school aims to use Foundry to scale the school’s offerings far beyond the 900 or so MBA students it admits each year. In that way, it is something of a case study for companies facing a similar problem: Consulting firms like PWC, law firms and, at least in one case, a church, are also experimenting with how AI can expand their services far beyond their human bandwidth. So are some movie directors – an AI version of the actor Val Kilmer, who passed away in 2025, recently appeared in the historical drama As Deep as the Grave.
When the team at Harvard started working on the AI project in 2024, they imagined a simple chat window that would be similar to ChatGPT, but equipped to advise entrepreneurs, said Katharina Rings, the project director of Foundry. But after the school released a trial product in October, she said, it heard from students that they wanted a more guided experience, which led to the current version, available since April.
Harvard Business School famously grades on a forced curve, and Foundry had to intentionally teach the AI to not be obsequious, a common trait of large language models, said Tom Eisenmann, a faculty co-chair for the project.
He said he’s made a point of letting participants know that the AI agents are programmed to be so tough that virtually no startup ideas posed to them have received a blessing on the first take. “We literally would have somebody spend 10 hours with the thing, just sort of trying and trying,” he said. “That’s wonderful, but really it’s not the point.”
The AI agents that chat with participants have been trained on the work of specific Harvard Business School faculty, but they also draw from wider knowledge when appropriate.
Foundry created the AI-generated videos of professors by working closely with the AI avatar maker HeyGen. The positive reaction to both the AI chat versions of professors and, especially, the AI video avatars came as a bit of a surprise.
Rings said that in tests, simply adding a photo of a professor to an AI chatbot increased engagement. “It was just so clear from what the users were telling us and what the numbers were saying,” she said. “It seems like people actually much prefer the AIs, and they actually trust more, if it is based on one specific person’s input.”
While anyone can apply to buy the Foundry course, many participants have found it through university departments that help students translate their companies into commercial businesses. About 45% of founders in the boot camp’s sessions over the summer paid individually. The rest were sponsored by universities, either through a paid subscription or a free trial that HBS Foundry offered to some schools. About 760 founders in the summer sessions found the program through their university.
Harold Solomon, a principal at Georgia Tech’s entrepreneurship program Quadrant-i, said the advantage for schools is they can serve more founders. “I can have 20, 30 meetings a week, but beyond that, I literally can’t,” said Solomon. “We can’t service the class, let alone the campus, in person constantly. So we need a way to scale.”
He incorporated Foundry into a 12-week entrepreneurship workshop for PhD students and faculty over the summer and signed on to continue offering the program to graduate students.
Foundry participants said that the unconventional AI instruction didn’t faze them. Some had already been asking AI for advice, and found the HBS version more tailored to their needs.
Pranav Sultania, a doctoral candidate in mechanical engineering at Boston University, has spent years developing an electromagnetic jet propulsion and pumping system for underwater applications. But he was only a few months into creating a business around it, called Magnesis Robotics, when the BU entrepreneurship centre gave him access to HBS Foundry.
Talking strategy with the AI avatars helped him focus his pitch. “As an engineer you want to think, Oh there’s so many use cases, surely when I pitch that to an investor that is more business worthy,” he said. The AI avatars pushed him to pick one to highlight.
It was also helpful, Sultania said, that the AI avatars were always available when he had time, which as a busy founder and PhD student, is not always predictable. “I kind of want the platform there that I can talk to, even if it’s a clone, anytime I want,” he said.
Eventually, Eisenmann said, Harvard Business School may allow other universities to add avatars of their own professors.
Though for all of the AI assistance, he acknowledged, one of the most popular features of the product has been the forums that allow entrepreneurs enrolled in the project to talk with each other: “They love the AI. They’re using the AI, they’re learning a lot,” he said. “But it keeps coming back to humanity.” – ©2026 The New York Times Company
This article originally appeared in The New York Times.
