The Exchanges

Q Really interesting. Okay. Let's do a couple of these examples here. HubSpot, VSP, just to be clear, are these real, these are real paying customers, real examples?

A They are real enterprise paying customers with logos that we are allowed to talk about. So these are not, uh, experimental budgets. Most of these customers are paying the cost of a human, um, plus and have multiple like suit, like HubSpot, for example, we are working on, I believe it is their fourth superhuman, um, to go live across their journey. Fiona, the one you're seeing here, was the first superhuman that we built for them. It's in the top left corner. Fiona is set to talk to their SMB business, so when somebody asks for a demo, they engage with Fiona, she gives the demo, she qualifies, she takes them through to close, and they were able to increase their revenue by 25% in their SMB small business segment.

AI assessment note: “They are real enterprise paying customers with logos that we are allowed to talk about.”

Q If someone's listening to this going, man, should I add this to my website? I get 10,000 hits per month, right? What would you tell them? Should you say expect that 20% of random cold hits to your website will engage with Mindy? Or what do you think the hit rate is?

A Yeah, I mean, think about it. Most websites have some kind of chatbot on it today. Like, what are you getting in that chatbot today? Today, you're probably getting maybe two to five percent. You can at least double, if not five X, what you're getting off your traditional chatbot with a superhuman. And our goal isn't just to be in the real estate of the chatbot. We live our, we eat our own dog food or drink our own champagne. And the fact that Mindy is like front and center on the website, right? Like she is all you, that's all you get. However, our customers have it as a chatbot. They have it as an embedded experience, like as a banner on the homepage of some customers who have it in the top navigation bar, like Talk to our superhuman. So we have it embedded in outbound campaigns. Let me send you a campaign and have a highly dynamic, personalized conversation with that user and offer it instant versus booking a meeting or downloading a white paper, right? So she lives all over the place. She lives on LinkedIn profiles, et cetera.

AI assessment note: “You can at least double, if not five X, what you're getting”

Q website and she has a thousand conversations and I was like, oh my gosh, I thought she was going to have five. And now I owe Amanda like a million dollars when I thought it was going to be like a hundred K that makes it really hard for me to predict like my budgets for what my CFO is that, did you experience that firsthand? Is that what happened?

A No, we, we actually said you would buy a bucket of conversations. So let's say you, we would estimate, okay, you're going to have a thousand or 2000 conversations a month. So let's call that 12,000 or 24,000 conversations a year. And I would sell you a bucket of those conversations. And now what we're moving towards is we still have that concept, but I'm basically giving you exponential what we think you're going to do. Like if you have 10,000 visitors to your site, I think I get 20 to 40% will talk to the superhuman. If you got all a hundred percent, I'm not going to charge you more. So we're really looking at this of like, we want to be in partnership. We don't want To prohibit you and slow you down from using the superhuman. We don't want you to put her in a corner. We want to give her as much exposure, he or she, or they, whatever pronoun you want to give your superhuman, um, as much exposure as possible to drive as much value in your business as possible. So we're just trying to align back to the values of our customers. Um, and so it's kind of taking that away in a way that says, all right, I'm going to five X, I can't go to a hundred X cause then I'll be out of business, but I'm going to go to five X what we think you're going to use.

AI assessment note: “No, we, we actually said you would buy a bucket of conversations.”

Q quantified, I mean, everyone watching my show totally understands the old sort of SDR, BDR to AE ratios, and then there's CSM after that. Everyone understands that flow. How are you, I mean, you probably ran that playbook at six cents, actually, as you scaled six cents. How is this radically changing that? Is this just like a Is it a cost structure change? Is it something else I'm missing?

A If you look at the jobs report, AI has put on more jobs than has taken away. I think there are some legacy ways that we ran go to market that is highly inefficient, as you mentioned, but really a lot of companies today are stuck and they're not growing and they need to cut costs. They need to do two things. They grow and they need to cut costs. And so what bigger problem to solve than those two things through the form of an AI superhuman that basically can do the job and do it exponentially better, more efficiently, and Serves the buyer. So my ultimate goal, why I'm helping companies grow, I'm helping them cut costs. My number one thing that we are focused on is creating a better buying experience. The handoff from an SDR to an AE to a sales engineer to a CSM is atrocious. If you think about, and there's a loss of context and memory, we ask the buyers the same thing. We take them through the journey. And it's, if you put yourself in the buyer's shoes, the process that we go through with humans today is ungodly. So this is a new way in a new world to have unlimited memory, recall, capacity limits, can talk to any industry or vertical. Like for example, if you're talking to like a Cisco or a Databricks that sell to every industry and every vertical and anyone who needs data, for them to staff sales engineers and solutions engineers that understand the nuances of every vertical an…

AI assessment note: “The handoff from an SDR to an AE to a sales engineer to a CSM is atrocious.”

Q It's basically him. If you didn't know that this was AI, you'd think, well, maybe this actually is actually sort of Adam. I mean, Amanda, do you think we're working to a world where me and you are actually created, and it's okay that users are talking to a digital version of us, even though it's not really us? I mean, where, where those lines blur over time, you think?

A Yeah, we're not in the business of cloning, but we do. So we cloned AI Adam because he has a brand, right? So he has, like, another one is Jack, Jocko from Winning by Design, if you know him, but he has a very strong personal brand. And so he decided to clone himself as well. So we will clone, you know, if it makes sense for the brand of the, for our customer. Most of our, actually, an interesting one is Alteryx, is they took the founding, uh, the founder's wife, and then they made an altered version of her. So it's a likeness of her, which I thought was great. It was like a nod and a head nod to her without actually cloning her, per se.

AI assessment note: “So we will clone, you know, if it makes sense for the brand”

Q I mean, I'm using it here on my screen. I had to mute it because she started talking to me, but you guys get the sense of what's happening over here. Amanda, just out of curiosity, what's the tech stack here? I mean, is this like a HeyGen real time API plus 11 Labs voice on top of it? Or what, what's the tech stack here powering her?

A Uh, we use a multiple of different, uh, technologies. Some of it's our own, and some of it's, uh, third-party vendors. We actually have our own pipeline for most of it. We use multiple different face, uh, vendors. The face, I believe, will be commoditized at some point, so we did start out building face, and then we kind of said, you know what, we're more about the brain. We want to build the context graph and be the decision logic and the engine behind, so that she can give the real live demo, so that she can, um, support and create ROI calculators on the fly, etc. So, yeah, it's a combination of different things, and a lot of it's our own.

AI assessment note: “We use a multiple of different, uh, technologies. Some of it's our own”

Q Six cents. You then went out and I think you're launching here with like a massive raise versus I'm thinking if she really believes in this, why not use her own money and just fund this herself? So two questions there. Did you get really rich at six cents and why not sell funds now?

A I get really rich at six cents. Um, I think the story that I share openly, um, is yes, I left at a time when we were raising another round. It actually just happened to be that way. It wasn't like when I was leaving, we were timing it and planning it. I took five years off so you can do the math. It was fine. Um, Didn't plan on coming back to work. I always said I was done, and I was on the bench, and I was having kids, and I was happy. Uh, played a lot of golf. Realized that being a mom is the single hardest job in the world, so I decided to outsource that. Come back to work. Um, and did I sell fun? I mean, maybe I put money in. Maybe I didn't. I don't really share that. I don't think it's right.

AI assessment note: “Maybe I put money in. Maybe I didn't. I don't really share that.”

Q Can I ask you what you're doing today or a range?

A You can't. Let's just say we raised Very soon after. We've been in market for 18 months, so we've, you know, we've come up on a full year of renewal cycles. We had a 211% NDR, which is amazing. So these aren't experimental budgets. Almost every customer within 90 days of going live adds a second superhuman. So we have actually more opportunities when you were talking about those higher customers within our customer base than net new. I think a lot of net new are afraid. They're like, wait, can this do this? Is it going to hallucinate? Are we sure this can represent our brand and our product and this is how we want to go to market? But it's absolutely Absolutely wild when they see the results and they see that this works. And I can tell you, 18 months of being live, I have not had one single customer tell me, touch wood, that this doesn't happen, that the superhuman has hallucinated in a way that has negatively impacted their business. The only time we've ever heard somebody say that it said something they didn't want it to say is because they gave us content that was outdated. And so I was like, they point to it and like, wait, we don't do that anymore. And I was like, well, let's go look at what you gave us. Then we cite back to the source and like, oh shit, yeah, we used to say that. Okay, so she wasn't hallucinating. She was Actually using the content.

AI assessment note: “You can't. Let's just say we raised Very soon after.”