It’s 11:40 p.m. You’ve got a slide deck with a hockey-stick chart you don’t fully believe in, a lukewarm coffee, and a browser tab open to your practice board meeting. On screen, a Harvard Business School faculty member leans in and asks about your customer acquisition cost. You answer. The professor nods, gives feedback, waits for your next move.
Except the professor isn’t there. It’s an AI avatar — a rendered replica of a real HBS instructor, built into the school’s HBS Foundry accelerator program. The price of admission is $699.
I’ve reviewed a lot of AI tools that promise to replace something expensive with something scalable. This one is interesting less for what the technology does and more for what it admits out loud.
What’s actually on offer
The verified details are narrow, so let’s stick to them. HBS Foundry participants pay $699 and practice pitches and board meetings against AI-generated replicas of instructors. The avatars provide feedback during those sessions. That’s the product.
Note what that is and isn’t. It isn’t a professor watching you sweat and then telling you the thing you didn’t want to hear. It’s a simulation loop — reps against a synthetic counterpart wearing a real person’s face and, presumably, a distilled version of their teaching approach.
Why this is smarter than it sounds
My honest first reaction was cynical. My second reaction was that pitch practice might be one of the few use cases where a synthetic instructor genuinely earns its keep.
Pitch coaching is repetitive by nature. Founders need dozens of reps, not three. They need to answer the same objection about unit economics until the answer stops sounding rehearsed and starts sounding true. Human faculty time is the scarcest resource in any business school, and burning it on rep number seventeen of a seed pitch is a poor use of a very expensive person.
Simulated reps also lower the social cost of being bad. A lot of founders never practice properly because practicing in front of someone whose opinion they care about is humiliating. An avatar you can fumble in front of at midnight, repeatedly, with no reputational damage, removes a real barrier. That’s a genuine benefit, not marketing copy.
Where I’d push back hard
The problem with a replica is that it’s a model of a person’s outputs, not their judgment. A good instructor’s value shows up in the unpredictable moment — the follow-up question you didn’t prepare for, the observation that your entire framing is wrong, the decision to stop the pitch two minutes in because something more fundamental is broken.
That kind of intervention comes from pattern recognition across hundreds of real founders and a willingness to be blunt in ways a product built for paying customers usually isn’t. I’d want to know how the avatars handle a founder whose idea simply doesn’t work. Do they say so? Systems designed to give feedback tend to give feedback, which is not the same thing as telling someone to stop.
There’s also the branding question, and it’s the one worth sitting with. What you’re buying at $699 is proximity to the Harvard name delivered by a synthetic stand-in. The school’s reputation was built on human faculty, small rooms, and the discomfort of being questioned by someone who has seen your mistake before. Licensing that reputation to a rendered version of itself is a trade. Maybe a reasonable one. But it’s a trade, and institutions that make it repeatedly tend to find the underlying asset worth slightly less each time.
How I’d judge it if I were paying
Practical framing for anyone considering it:
- Treat it as a gym, not a mentor. Reps and muscle memory, yes. Strategic direction on whether your company should exist, no.
- Check whether the feedback ever tells you no. If every session ends encouraging, the signal is worthless.
- Compare against free. You can run pitch practice against a general-purpose model today for the cost of a subscription. The $699 buys structure, curriculum, and the crest. Decide whether those three things are worth the delta for you.
- Value the accelerator, not the avatar. The surrounding program, cohort, and materials are probably where the actual value sits. The avatars are the delivery mechanism.
The part nobody wants to say
Harvard shipping AI replicas of its own faculty is a signal about where paid education is heading. Not because a business school found a new revenue line, but because one of the most prestige-protective institutions in the world decided that synthetic delivery was acceptable for a piece of its teaching. Everyone downstream now has permission.
Whether this ends up as a useful practice tool with an honest price or the first step toward faculty-as-content depends entirely on how candid those avatars are willing to be. A simulation that tells you what you want to hear at midnight isn’t training. It’s a very well-credentialed mirror.
🕒 Published: