The AI Pricing Puzzle: No One’s Cracked It, But We’re Getting Closer

I've been in some board meetings where the person comes in, you're like this person does not want to be here and if it really isn't a fit as a founder, you should go back to the firm and be like, hey, this person either doesn't show up or doesn't seem that interested. I'd rather take a more junior person as a learning experience who's excited about this than a more experienced person who doesn't care. I think the biggest difference between 2024 and now is we actually have live contact. The old Mike Tyson analogy about everybody having a plan to get punched enough.

And 12 months ago it looked like 29% we're leveraging a flat fee subscription pricing and 20% were doing hybrid. Fast forward to today, 6% are continuing with that flat fee model and 41% are leveraging hybrid pricing models.

Welcome to the Learning Corner by Precursor. I'm Mia Farnham and I'm joined by our GP and Managing Partner Charles Hudson, where we will be talking through the articles and goodreads that caught our team's ATT over the past week. We took a break on the POD over a Fourth of July week and we're really excited to jump into this week's topic. But before we do, I did want to share a shameless plug about our upcoming pre seed bootcamp that Precursor Ventures is hosting on September 25th in New York City. It's a full day of programming for people who are maybe early on in their startup journey or thinking about starting a company and we bring in early stage investors in our network, founders across our own portfolio, and incredible lawyers who've made their entire careers supporting startups and a lot more. And it's application based and we'll be launching that at the end of the week on our website and social channels. So definitely be on the lookout for that.

It's one of my favorite things that we do. This will be the third installment but the first one we've done in New York and I really enjoy these because it's a chance to spend a bunch of time with people and get to know them in a non transactional, non pitch format which for me is very hard to do. So so a lot of the alums of our bootcamps in the past have become friends. Some of them we funded, some of them we've helped get funded by other people and so I'm excited to do this in New York.

Be on the lookout for that. Let's get into it. Following the news that Eric Paley, co founder of Founder Collective has been appointed Secretary of the Executive Office of Economic Development for Massachusetts. Our first topic is written by Eric's longtime partner Micah Rosenbloom, who shared a thoughtful reflection on a frequently occurring dynamic in venture it's the the Adopting Parent Investor When a key investor or board member transitions off a company, whether for retirement, a new role or life changes, it can leave founders in a vulnerable spot. Maika unpacks how stepping into an inherited investor relationship isn't just about capital, it's about continuity, trust, and honoring the history that's already been built. He offers principles for doing this well, including being upfront about your newness, committing fully to the role, so cutting down on revolving doors and avoiding the trap of trying to be a one to one replacement. If you are in venture, you know this adopting parent investor concept happens frequently. Having seen boards across 450 plus companies across 5 funds Charles curious what your thoughts are here.

After reading it, I read this and I thought it was the best articulation of this that I've seen to date and I've experienced it from literally every dimension possible. When I had my last startup, we had a partner transition out of the firm that had led our seed round and we got replaced with someone else who is a friend who I love and adore but knew nothing about what we were building. Like we were building a games company and games was not this person's domain of expertise. The person had never been to a board meeting and he did his very best to be a champion and an advocate. So I've seen it from a founder experience. We've had some folks depart precursor and I've inherited some of their portfolio companies and have tried to figure out a way to be helpful with those companies without again to Micah's point, trying to replicate who the people are who made the investment because they're them and I'm me. And I've also been on boards where like a new person shows up from another firm and you're like well how's this going to work? And I think probably in the most challenging of all, a very close friend of mine who I was in with passed away and one of his partners had to step in. And I think this is becoming a much more common issue in venture because it used to be people didn't really leave firms and people didn't get fired, pushed out, asked to leave, decide to leave, whatever. There just wasn't that much turnover and now we're in a world where I think there's just a lot more turnover in venture and you know, rules I used to give people would just like, oh, if you have the managing partner and founder of the firm on your board, you're probably safe. That person's not going anywhere. As we just saw with nfdg, even the managing partners of the firm can leave. And with some of the firm breakups we've seen, that's not safe. So one thing I've noticed is a lot of the firms that plan to be around for a long time and have the capacity to do so, I think have started staffing a one plus one model where there's a primary person on the board or primary point person, but there's somebody else who's the number two chair who's around and familiar with the company, so that if anything were to happen to the primary person, there is in fact somebody up to speed. This is really hard to do, though. If you're a small seed firm and you're maxed out on board capacity and maybe you only have two partners, you're like, look, we can't have every partner and every deal. But I do think it's more important than ever to think about what you're going to do. And I think the one thing that Micah said of all the things that I think is really important is don't try to be the person you replaced. Just come in and be yourself. And also just recognize that it's probably a big deal for the founder to lose their advocate. I have a founder in our portfolio where the person, she chose a firm literally to work with, one very specific partner, and that partner left and the founder was devastated and said, what do I do? I said, well, you should go to that firm and pick whoever you're most excited about working with, because I'm sure they want to keep you happy. And this is a window for you to say, this is what would keep me happy. And she ended up picking somebody who she's been really happy with. But I think this is going to become a recurring problem. I have a feeling this is a post that people are going to reference.

Pretty often and especially in this time of increase on secondaries, where maybe your switch off the board is happening a little bit sooner. I feel like for early stage investors, especially when you come super, super early, you come off boards or you join in on companies before boards are even created, but the fact that you could possibly jump out of the journey earlier on than maybe founders had anticipated is a new one entirely. But it doesn't break the relationship in a 100% capacity level. And I feel like you can speak to that being a first check in for companies that are now at series D and have exited and beyond.

Also, we have people who've left their firm but are still close to the company and still provide some level. It's interesting, when I left uncorked, I had a handful of companies where I was still their primary contact that they didn't need very much help. But just from a continuity of relationship standpoint, I think that's the best of all worlds. If people have the capacity and obviously if you've taken a political office, there's not really any way to be a venture partner to the side that probably doesn't work. But I think wherever possible I always try to think what's in the best interest of the entrepreneur. And oftentimes it's some ramp down with your original person while you're ramping up with your new person is the best if you can swing it, but you can't always swing it right.

And then along with not trying to be the person that came before you from a one to one ratio, I think Maika said another valuable piece of advice was disclaimer your newness rather than pretending expertise. I know that a lot of investors in this space that's kind of like their favorite thing to do is pretend expertise, but I actually find in these scenarios it's the most harmful thing that you could do. I think we've all been on it on the other side when an investor comes in and just blurts all these things that they assume about your business without having built that relationship prior. And that's actually quite harmful and disruptive for founders.

And I think not all of the catch up needs to happen in the board meeting. But I do think whenever I've been new I've been like, how can I show up to the board meeting with the most context? Sometimes looking at the old board materials is helpful. Sometimes talking to the current board members or outgoing like, hey, what are the big issues at the company? Oh, the VP of marketing, VP of sales, or so you have some context. But look, it's inevitable that there will be things that the board has discussed before that will strike you for the first time. And I've always used a disclaimer. It's entirely possible that you guys have talked about this before and I assure you over time there'll be fewer of these. But why are we doing X, Y and Z thing and not a, B and C thing? And I've always been surprised. Sometimes people are like, well you know, we did talk about that two years ago, but it's probably worth revisiting. Yeah, because things have changed or the world's different or here's some new piece of information that we didn't have at the time. And so I do think sometimes being new with fresh eyes is actually helpful and allows you to provide context in a way that someone who's been doing it for a while won't or can't.

And I find for even folks who maybe are the more junior replacement for somebody who came before you on a board, the most valuable piece of advice that I think I got from another managing partner at a different fund was the best set of skills you can develop early on in those boards is being like a very active listener. You don't need to make a splash and create a whole host of seven different ideas that you have for the roadmap of the company.

I agree.

Any recommendations, Charles, for founders on navigating this that you already haven't provided?

I do think it's always good to find out who's going to be the person taking over two things. What's their baseline knowledge of the business? And candidly, what's their baseline level of interest? Not to use a curb your enthusiasm reference, but sometimes these things get foisted on people and it's not fun. When I've been in some board meetings where the person comes in, you're like, this person does not want to be here. This person drew the straw that they had to take this board. And like, their body language from day one is such that my firm is making me do this. And if it really isn't a fit, as a founder, you should go back to the firm and be like, hey, this person either doesn't show up or doesn't seem that interested. I'd rather take a more junior person as a learning experience who's excited about this than a more experienced person who doesn't care. And so I'd say if the first match really is a misfire, don't be shy about going back to the firm.

Also, everyone else on the board who's remaining there probably knows that too.

Yeah.

Our next topic is a recent episode from the Peel with Turner Novak featuring Samir Kaji, co founder and CEO of Allocate, where they have a conversation for more than an hour about all types of happenings across the venture landscape. They break down the current shifts across the VC industry, walk through the importance of entry price in this era of secondaries, advancements provided by AI that SaaS companies should have a strong pulse on, and a lot more. They talked a little bit about Allocate, which is the company Samir created to transform the private market investing experience by building building a curated managed marketplace for financial advisors to have exposure and access to private market investing opportunities. Samir in Turner talk about what Allocate got right on the platform that other private market aggregators have gotten wrong. So wanted your perspective on that from a fund manager perspective.

First, I'm a huge fan of Samir and Hannah as they know. Samir was, I think, the first person to ever make meaningful family office intros to me when I was starting Fund one. I remember, you know, RIP First Republic, but I rem going to his office at First Republic and sitting down with him and going over my deck. And his first ask wasn't let me set you up with a bank account. His first ask was let me get you in front of a few family offices who I think would be interested in your pitch. And he's always had this spirit of giving in a way that it's beneficial for both the person he's helping and the person he's introducing. So I hope he and Hannah and the team continue to be a positive force for the ecosystem. I really like this episode. I don't always listen to things north of an hour. It's a big ask, but in this case I really did feel like like there were a lot of really important things that got discussed. One was if you're a fund manager and you're struggling with like positioning and marketing your fund, I really encourage you to listen to the middle chunk of this episode because a lot of the wisdom that Samir drops in this episode is about how to position and market yourself as a fund manager. And ironically, some of it is directly saying do this, don't do this. But some of it's also just talking about how they positioned Allocate. And I think one of the big things they understood was to have a product that works for the RIA and Family Office channel. Need really awesome funds on the platform. But you also have to think about what are the problems that those folks are trying to solve. And as somebody who's begrudgingly or maybe accidentally become an LP in over 40 funds, including mine, I have 50 different portal logins I have to go to to find my K1s and financial. There are all these challenges that they've solved, but I think they worked backwards from if we were going to make this available to many more people, what would we have to solve? I think the other thing that I really liked about this episode is something that, you know, we're doing internally. Samir is playing with a lot of AI tools directly him. And I just, I think everybody should be doing this because otherwise one of two things can happen. You can become overly cynical because you use ChatGPT once and it hallucinated and you're like, none of this stuff works. Or you just read the headlines and you're like, oh, AI is perfect and beautiful and magical and I'm not going to have a job in a while. And I think the answer is most of this stuff works well enough for you to see that it's going to be a big deal in the future, but a lot of it still requires supervision. And I just thought Samir, I think, knows maybe more about the world of FOS and RIAs and how, how they're thinking about venture as an asset class. And there's just a ton of wisdom in, you know, everything from how do those folks really think about big firms, how do they think about portfolio construction? And then the thing that's hard to get, how do the clients of the people who are on allocate think about venture? And so I just thought it was a really great conversation and if you have time, it's well worth a listen. And he's got a lot of really good insights.

Going back to your AI comment, I think a lot of my LP friends are maybe a little bit wary about using AI tools themselves or, or spending time learning about them. And I think that's the most valuable thing you can do because I assume that when you're getting pitched by fund managers, they're going to tell you about all these different bespoke AI tooling that they've built across the fund. And it all sounds really wonderful and unique, but if you take a look under the hood and if you did a little bit more research on your own, you would know that it's quite similar to what five other funds down the street are doing. And so I think that's the biggest benefit of actually making sure that you're getting smart on those things as well as the space is evolving.

I think so. And I think the other thing that, that I'll say is I do think one part of that interview I really like Samir, is like, look, I think a lot about, well, what are the many roles that Cambridge Associates fulfills? One is vetting fund managers on behalf of other people. And having spent time with Samir, I really think they have a point of view that I agree with, which is they want people to be successful in venture they want people to put money into the category and get money back. But to do that you have to invest in good funds and somebody's got to help you figure out what the good funds look like. And my fear is always the, that we'll meet family offices and they'll tell me they're in a bunch of funds. And I'm like, well, I know none of those funds have performed. And that is going to color your viewpoint on whether a venture is a good asset class. And so I think having somebody who can also be a bit of an arbiter of quality is really important.

I talk about this a lot of just having somebody who could kind of bridge and share a little bit more color on what's out there in the market for LPs but help make that decision, I think is something that we're meeting across the ecosystem. A conversation thread that they both had was on the importance of timing the market from a selling standpoint. For early stage investors looking to get liquidity from a second secondary opportunity. A really good quote I heard was something along the lines of you can't control the exit price, but you absolutely can control the entry price. So the price that you come in on from an initial check writing standpoint, how do we marry that concept and discipline with sky high valuations of companies raising money across AI today? And I know we've had a lot of internal conversations around like now with this AI talent war. I think a lot of interesting investments that maybe aren't as wonderful as they would have thought they would be are going to get bailed out because there are people who are willing, willing to put money to buy the talent themselves. So like, how do you detangle all of this?

A long time ago someone told me not every good company is a good investment. And so I think about that a lot. I think the other thing that's it depends on where you land. You know, we talked before about the VenCap piece, which is like, hey, if you really believe we're going to a world where exit scale for the winners is an order of magnitude or 2 larger than what we previously believed. You can pay an order of magnitude or 2, maybe not 3, as we're seeing in some cases on the entry price, provided that you get into winner. And I talked to one LP who was like, look, what matters is the scale of the winners. If you lose your money on a $50 million entry price or on a 5 million dollar entry price, it doesn't matter to me because you still lost your money unless somehow the company sold in such a way that you would have made money at 5 but you couldn't at 50. But I just think you can't use the pricing outcomes for the outliers to determine the average. And so I think what's happening now is I see companies that are getting priced in such a way where I'm like as a seed fund, okay, that's a hundred million dollar post. So that needs to be billion company to make the underwriting work for you. The $10 billion companies certainly exist, but there's not that many of them in the private markets and the ones that are supported by, for the most part by real revenue and product. And a lot of these companies are being assigned very high valuations with very little evidence of what they're going to build on the other side. And so I find it hard under our model we're a sub $100 million fund. I think entry price really matters and you know, there's 10 or 15% of the time where we're going to stretch and do something at north of 20 and south of 50. But I wouldn' to build our entire fund with those enterprise and at the same time, yeah, I have friends who that's all they do. They're like I just want things at 30 to 50 and I want to buy 10%. And I think these are truly exceptional special companies that are going to be worth billions or tens of billions of dollars. And I just think it's a different, that's a very different way to invest.

Well, our last topic is a piece published by Emergence Capital co written by Jake Sabre, Josh Bloom and Madhavan Ramanujam. Discussing the art of pricing AI software. The piece delves into the most frequently asked question today, how should I my AI product? We've talked about the many different models on the podcast see Usage Outcome based Hybrid. And throughout the piece, the authors do a really great job of outlining what has changed about pricing in this evolving AI landscape. The piece highlights three fundamental shifts that are affecting software pricing as it pertains to AI 1 AI enables true outcomes based pricing, 2 AI taps labor budgets and 3 LLM usage adds a completely new cost dynamic. There are great tips for founders on how to begin exploring how much to charge charge for your AI product, along with great summaries on the current and future state of AI pricing. Ultimately, the piece concludes with the quote, the companies that define the AI era will be those that align their pricing with the value they create. We've had a few different conversations about AI pricing on the podcast. So I wanted to get your thoughts here first. And also what are some palpable differences in talking points here that hadn't existed maybe in 2024. About the same topic.

I really like this piece because I thought it was something I could hand a founder and say, here's a framework you should figure out of business you're building and this will give you some tips as to maybe what's possible for you. I think the biggest difference between 2024 and now is we actually have live contact. The old Mike Tyson analogy about everybody having a plan to get punched in the mouth. Everyone's like, oh, customers are going to want usage based pricing or customers are going to want this or sell the service, not the software. These were largely theoretical arguments in 2024 and I think now the answer to me seems to be it depends. As in it depends on the service you're providing, it depends on what your customer wants to buy. It also depends on is this an experimen budget or is this a production real budget. And my sense is we don't know how this is all going to play out because this is all happening in a world where customers have historically had some aversion to usage based pricing because it's hard to budget. And also the more of the product you use, the more you pay. There's some things like payments where it's just, it's embedded. And I think we also have a legacy of per seat pricing for SaaS and a bunch of vendors who are sort of very familiar with that kind of pricing. And then I think as the emergence folks mentioned, the LLM usage is this like drag which is a different element. CL allowed the marginal cost was was pretty low. And with LLMs today at least the bill can get quite large. You know, if you're really banging on cursor, the bill can get big. My hunch is that this is going to continue to evolve. I think some of these things tapping labor budgets sound great, but sometimes the person who has the labor budget and the person who have the software budget are not the same person. And so how do you even deal with the internal silos at companies? And the outcomes thing to me is very interesting. But what's an outcome worth depends on depends seems to me to depend a lot on the company and their customers. So to me I write this as this is a very good framework for figuring out a starting point if you're a software provider. But we're all going to have to learn together over the next couple of years and I wouldn't be surprised if there is no dominant model. But if in the end we end up with for different products, there's a slightly different way that people pay for it. But within product categories there's some standardization.

Yeah, it's really clear right now when they shared a really great graphic about where AI product pricing was 12 months ago and where it is today. Clearly we're still mixed. I think 12 months ago it looked like 29% we're leveraging a flat fee subscription pricing and 27% we're doing hybrid. Fast forward to today, 6% are continuing with that flat fee model and 41% are leveraging hybrid pricing models. So to your point, we still don't know, but we know it's a mix or we're at least that's what we're testing out.

And I'm guessing that a bunch of those people who gave people a flat fee realized they were paying more for compute than they were getting in revenue. We can't let people just have either unlimited or we'd set the tiers wrong. And this was actually costing us way too much to run this service and we kind of can't allow that to happen.

Last question for you. They walk through using the acceptable expensive and prohibitively expensive framework to determine where willingness to pay lands for AI products. One of our portcos Superhuman leveraged this technique to price the product, which obviously has been successful. They just had a recent exit over the weekend. How does this technique take form when perhaps customers don't really have a gauge as to what they're actually buying? I know we harp on this a lot, but I'm finding that customers on the enterprise side are oftentimes disappointed with the ROI for some of their AI product purchases, especially over the past year. So how do we continuously evolve in educating them and pricing correctly?

I seem to find most AI products I encounter today, this is a broad generalization, are either dramatically underpriced or way overpriced. There's a couple things I use. I use a transcription product called Whisper. I'm not an investor. Whisper feels like maybe about right. I just think the intuition here is good because I think most startup founders, if I were to generalize under price their product because they don't want price to be a reason that early customers say no and they end up making the product too cheap. And if you just say instead we're going to anchor our pricing on something that feels like it will cause the person to pause and think about it, but where they will still ultimately say yes, that's a better instinct.

Yeah, like not being able to test the willingness to pay. And I have a list of them in my head and I'm not going to name them on the podcast, but like I'm using them all for literally $0, if not less, $9 a month. And it has not changed since I started using it nine months ago. And also I think the problem is if you don't change the price in a specific period of time, there are other tools that have jumped into the space also. So instead of me deciding to increase what I pay for the product, maybe I will just go for something else.

I think it's interesting that ChatGPT is 20 bucks, but so is Claude. ChatGPT Pro is 200. You know, sometimes the person who goes first sort of sets the reference price. By the way, the pricing work that the 49 Palms folks have done is really good for people who are looking for tips. I think those guys are excellent. But ChatGPT has told me that general purpose access to an LLM for a base paid level should be around 20 bucks.

Yeah.

Why? Why 20 bucks? Because they made it 20 bucks. I don't know because they made it 20 bucks. If they made it 10 bucks, that'd be my reference price. And now someone wanted me to pay 50 bucks for the LLM, I'd be like, well, I mean, ChatGPT and Anthropic are both 20 bucks a month. Like, why are you two and a half times as expensive? That doesn't make any sense.

It's all going to be handholding customers. Also on the enterprise side of educating them what's available that other folks don't offer. And that's what kind of warns the price increase. But I don't know how long that's going to take. Well, we are all out of time today. Thank Carl's for joining in. Thank you to our listeners for jumping in. Please like and subscribe and we will talk with you all next week. Thank you.

Bye bye everyone.