It’s funny how much we go through life without that same sort of feedback. We have dozens of interactions every day and a mental model of how we’re perceived. But the feedback loop is pretty slow and imprecise. I’ve often wished I could have a hidden camera follow me around for a few weeks and give me a more honest mirror of what it’s like to interface with me.

Recently, I got a taste of this unexpectedly as a guest in the Founder Launch class at HBS. The format was simple — I met with a bunch of companies over three days, evaluated them, and shared my notes on each one. But there was a twist. The founders also evaluated our interactions and provided feedback on what it was like to meet with me.

Then an LLM synthesized all of it, and the professor drew out the most interesting tidbits for discussion in class.

One thing he focused on was how I evaluate companies. The lesson for students was that different investors have different mental models, and he calls this out by comparing synthesized feedback between investor guests. This was interesting but not surprising. The model picked up on my core heuristic quickly: I zoom way in and then way out. “Is this something that might really work right now?” This means a lot of focus on the practical details of initial product and go-to-market, sort of skipping the business-schooly stuff like TAM analysis, the competitive matrix, the accumulating advantage. I then flip to “why does this matter?” What does the founder think is at stake if they win in this market, and what might the company become if everything goes right? My focus is founded on the idea that traction creates opportunity, and that the best founders are exceptional at the details of their business while also being great big-picture strategists.

But the more useful insight wasn’t about how I think. It was about how I’m perceived.

The Feedback I Didn’t Know I Needed

The LLM’s synthesis drew out one clear theme: I’m hard to read.

Not cold. Not unexcitable. Just… opaque. Founders couldn’t tell if I was enthusiastic, disinterested, or something in between. Most walked away feeling like we had a thoughtful conversation, but I was probably lukewarm.

This landed hard because I know exactly where it comes from. I have a huge pet peeve with VC/founder interactions. VCs have a well-earned reputation for “maintaining optionality” — smiling through meetings, saying encouraging things, giving founders the impression that things are going great. Then, days or weeks of ghosting followed by the non-specific pass email.

VCs also love to give advice, which founders often take as a sign of interest. But I strongly believe most VCs give terrible advice. It’s highly unlikely that a VC knows more about a business than the founder. And what they do know is likely biased and not nearly as transferrable as they believe. So I’m hesitant to advise on the specifics of a business except for the rare time (less than 10%) when I have extremely high conviction that I have something to offer that is non-obvious and accurate.

These are my two pet peeves, and I’ve developed behaviors to avoid both of them. But the feedback showed me that in trying to avoid one failure mode, I’d backed into another. My overcorrection left founders feeling like our interactions weren’t great — not because I was negative, but because they couldn’t read me at all. They walked away unsure. And uncertainty, it turns out, feels pretty similar to “that was a forgettable meeting.”

Different and Actionable

VC’s are generally starved for feedback. We might hear secondhand what founders say about us. Someone at YC might give us a sense for what is said about us on Bookface. But most of the feedback is general and unspecific.

This was very different in a couple cool ways. First, it was structured and synthesized, not anecdotal. It wasn’t one founder’s bad day — it was a pattern across multiple interactions, distilled by a model with no incentive to be polite or political. Too obvious to ignore.

Second and more importantly, it was fast and actionable. I got the synthesis after half my meetings, and I could quickly spot the pattern during the class discussion. The conversations were still fresh, and I could do something differently the very next day.

Day three was spent meeting with students in the second section, ending with class at the end of the day. I internalized the feedback and approached my interactions a little differently. I didn’t overhaul my personality or start performing enthusiasm I didn’t feel. But I made a few small adjustments, many of which were probably subconscious.

The feedback was clearly different. Founders didn’t suddenly think I was more positive about their businesses, but they generally felt like they knew where they stood with me. That led to a more positive impression of the interaction overall. This was true even for the company I had ranked as least likely for me to invest in. That founder actually remarked that the goals of our meeting were met and that we had a very positive interaction, even though it was clear I’d be a very unlikely investor in his company.

AI Mirrors

Most of the conversation about AI in venture is about using it to evaluate founders and companies. Better deal screening. Automated due diligence. Pattern-matching across thousands of pitches. Great stuff.

But I think an underrated use case is the opposite: using AI as a mirror to evaluate yourself. And this obviously isn’t just a VC thing. As software does more and more of the work, how effectively we interact as humans becomes more and more important, and AI gives us the tools to level up in this department.

I’ve seen a few products in this space already, and although some are sort of Black Mirror-esque, I’m generally bullish about their potential. Feel really isn’t real, and the path to growth involves accurate, specific feedback. This experience was eye-opening and is motivating me to give more of these tools a try. If others have experimented with anything like this, I’m all ears!