Six hundred and ninety-nine dollars. That’s what Harvard Business School is charging for its Foundry bootcamp, an eight-week program for entrepreneurs that now includes something I never thought I’d type in a review: AI avatars of actual Harvard instructors, built by HeyGen, giving you feedback on your practice pitches.
I review AI toolkits for a living, and I’ve seen a lot of avatar products come across my desk. Most of them get used for marketing videos and onboarding clips. Watching Harvard put this category of tool to work in an actual paid education product is a different kind of test entirely, and it’s worth walking through what this signals.
What You Actually Get
Based on what’s been reported, the Foundry program combines a few components:
- An eight-week bootcamp structure aimed at entrepreneurs
- Live sessions with instructors every week
- Virtual simulations
- AI avatars of HBS instructors that provide feedback during practice pitches and board meetings
The important detail is that the avatars aren’t replacing the humans. The live weekly sessions are still there. The avatars fill the gaps between them, which is the part I want to focus on, because that’s where this either works or falls apart.
Why the Practice Loop Matters
Anyone who has coached founders knows the real bottleneck isn’t information. Startup advice is everywhere and most of it is free. The bottleneck is reps. Practicing a pitch in front of someone who pushes back is expensive because that someone’s time is expensive, and a Harvard instructor’s time is about as expensive as it gets.
So the pitch here is essentially this: the human runs the live session, and the avatar absorbs the repetitive drilling. You can flub your board meeting simulation at 2 a.m. as many times as your ego can handle, and no professor has to sit through it. As a product design, that’s genuinely sensible. It puts the AI where AI is strongest, on volume and availability, and keeps humans where humans are strongest, on judgment and accountability.
The Questions I’d Ask Before Paying
Now the reviewer part of my brain kicks in, because a sensible design on paper doesn’t guarantee a good product in practice. Here’s what I’d want to know before handing over $699:
- Is the feedback specific or generic? Avatar tools live and die on the model behind the face. A photorealistic professor who responds with vague encouragement is a very expensive rubber duck. The value depends entirely on whether the feedback engages with your actual pitch.
- Does the likeness add anything? This is my honest skepticism about the HeyGen layer. If the underlying feedback is good, would it be any less good coming from a plain chat interface? The avatar of a real instructor adds presence and, frankly, brand weight. Whether it adds learning is an open question.
- How does it handle bad ideas? The hardest thing a real advisor does is tell you your idea doesn’t work. AI systems tend toward agreeableness. A practice-pitch tool that flatters you is worse than no tool at all.
What This Means for the Avatar Category
Whatever you think of the program itself, this is a meaningful moment for HeyGen and for avatar tools generally. Harvard Business School attaching its instructors’ faces to a synthetic product is a level of institutional buy-in this category hasn’t really had. Institutions like this are cautious by nature, and they’re putting real people’s likenesses, and a real brand, behind the output.
My read: this is the direction avatar tech had to go to matter. Talking-head video generation was always a feature looking for a job. Interactive practice and simulation, with a human program wrapped around it, is an actual job.
My Verdict, For Now
I haven’t run this program myself, so consider this an analysis rather than a hands-on review. But the structure passes my initial sniff test: humans for the live teaching, avatars for the unlimited reps, $699 for the package. That’s a coherent product, not a gimmick bolted onto a brochure.
The thing that will decide whether this is a template or a cautionary tale is the quality of the feedback coming out of those digital professors. If it’s sharp, this becomes the obvious model for professional
đź•’ Published: