Venture Everywhere Podcast: Mark Heynen with Noah Spirakus
Mark Heynen, co-founder and CPO of Knapsack, chats with Noah Spirakus, Partner at Everywhere Ventures.
In episode 60 of Venture Everywhere, host Noah Spirakus, Partner at Everywhere Ventures, talks with Mark Heynen, co-founder of Knapsack, a free, private meeting notetaker and automation platform. Mark shares his startup journey, from fintech to AI, driven by his passion for expanding access to essential tools. Mark also discusses how Knapsack leverages AI to protect data and build user-friendly applications for diverse customer bases, shaping the future of how businesses interact with technology.
In this episode, you will hear:
How Knapsack's AI automates meetings and manages contacts, enhancing productivity.
AI models are rapidly becoming commoditized.
The vulnerabilities of storing and processing data on third-party AI cloud platforms.
Underutilization of AI in financial services, particularly in relationship management.
Why companies delay AI adoption over privacy, regulation, and security risks.
If you liked this episode, please give us a rating wherever you found us. To learn more about our work, visit Everywhere.vc and subscribe to our Founders Everywhere Substack. You can also follow us on YouTube, LinkedIn and Twitter for regular updates and news.
TRANSCRIPT
00:00:00 VO: Everywhere Podcast Network.
00:00:14 Jenny Fielding: Hi, and welcome to the Everywhere Podcast. We're a global community of founders and operators who've come together to support the next generation of builders. So the premise of the podcast is just that, founders interviewing other founders about the trials and tribulations of building a company. Hope you enjoy the episode.
00:00:34 Noah: All right, everybody, welcome to another Venture Everywhere podcast. My name is Noah Spirakus, Partner at Everywhere Ventures, and I'm super excited today to welcome Mark Heynen. We were just talking about this beforehand. He is a fellow addict of the startup, and I'm excited to go through his background, what he's working on today.
00:00:51 Noah: But it's fun to find a commiserating soul who doesn't know what to do besides startups. I've gone through your background. It's been fascinating to learn, but I'm going to give a quick overview, kind of talk about that. But I'd love to hear more of the in-betweens and we'll talk about Knapsack and what you've been working on lately.
00:01:08 Noah: But Mark's background, built and sold several companies. The biggest through line that I found is you just love helping people that don't have access to things. It goes all the way back to feature phones and Facebook, but fintech tools, bringing people to AI, which I'm sure we're going to talk about a lot today, but super excited to have you, Mark.
00:01:30 Mark: Really happy to be here. Thank you very much.
00:01:32 Noah: I'll puff them up. That way you don't have to talk about yourself too much, but you've built and sold several different companies. What got you into startups? How did you get into this fascinating world?
00:01:42 Mark: Well, really appreciate it. Ultimately, it was by accident. I think this actually happens to a lot of people. I was in London. I had my internet analyst job in 1999, helping a large conglomerate in London figure out e-commerce. This is company called Kingfisher. If you combine Home Depot and Best Buy and put them in Europe, that's this company, Kingfisher.
00:02:03 Mark: The chairman, this older British guy, brought me into the office first day and asked, what is the internet and how do I buy it? And so very elementary understanding of how the internet works. He was convinced the Americans were behind it somehow.
00:02:16 Mark: And ultimately we went down this path of figuring out different e-commerce strategies. I took a couple of trips to Silicon Valley to help them navigate that. And I realized very quickly that there was a lot of analysis going on, not a lot of execution.
00:02:29 Mark: And I just found that I personally just wanted to get out and build as soon as possible. And it felt completely off, because what I saw working in California was building, and what I saw happening in London was analysis. And so I realized this about myself, that I really prefer the building process.
00:02:49 Mark: And I identified a business opportunity there, which is effectively price tracking for consumers and businesses. There was no one really doing that at the time. And so I left and started that company with a friend and went through that process of building that company and learned a lot about myself at the time. I was in my early 20s, but also learned that that is an activity I really enjoy doing.
00:03:10 Noah: What was your big takeaway? I know looking back on my first company, so many mistakes, so many scars that luckily I was able to bring forward into the next one, but what did you learn during that process? Early internet. Amazon, I think, at this time is only selling books. What did you learn from that process? Was that Electrobug?
00:03:28 Mark: It was Electrobug. That's right.
00:03:29 Noah: Okay.
00:03:30 Mark: So the first thing I learned is to choose your investors wisely. We did not choose our first investors wisely. They were the first ones willing to write a check. They did not end up being great investors and I ended up having to recruit an additional set of investors who were amazing to buy them out.
00:03:47 Mark: And that ended up being an arduous process. That second batch of investors ended up being amazing because I was able to meet Esther Dyson and a few really important angels who were amazing mentors to me during that very difficult time. But that was a very critical lesson.
00:04:03 Mark: I think the second, which is related, was it's really all about people first before product. If you assemble the right team around you, great things happen. And I did not realize that because I had this myth in my head that great product ideas just come to you and just by pure grit, you can bring those product ideas into the world and the customers will come to you if you bring those product ideas into the world and just find anybody around you to help you build those products.
00:04:28 Mark: And really, you're building a company as much as building a product. In fact, you're building a company first. And I actually learned that the hard way in my first few years and ended up building an amazing team. But it was a hard-won lesson.
00:04:49 Noah: Yeah, that makes perfect sense. I'm going to skip forward till today and we can go through the background also, but when we talk about people, we talk about trust is usually a big component of that.
00:04:51 Noah: And you talk about building products and we're in this age of AI. A friend of mine just texted me like, what's the AI coding tool you told me last month? And I was like, oh, it's a new one this month. Someone else has a new thing.
00:05:02 Noah: As you look at the ecosystem that you sit in, caring about people, caring about knowing that you actually need to solve the problem for your customer base and not just build a product and throw it out there, what's your general view of how AI, which I like to remind people that we used to call machine learning, what is the general space to you being in the middle of it? Are people actually solving problems? Is it still just wrappers on top of APIs these days? And how do you feel like you guys sit in the middle of that?
00:05:30 Mark: The thing I noticed very early on, and which I'm still noticing over and over again, is that the LLMs are being commoditized dramatically and very quickly. And we noticed this ourselves because we could swap different models very easily, initially running locally on the computer. We're doing private AI, but also in some of our inference providers.
00:05:53 Mark: And so there was very little switching cost. And that was something that I found was not being talked about very much. And I think now with DeepSeek coming out, people are realizing this is actually something that will be commoditized fairly quickly.
00:06:05 Mark: Then there's a second-order question, which is the wrappers. Are the wrappers very valuable and where does that value come from? And will OpenAI build all the wrappers and therefore capture all the value from all the startups? I think that's a really interesting question.
00:06:19 Mark: I think the application layer has a tremendous amount of value. There's a lot of work involved in just making this usable for people who are not technically as proficient. And you have a bubble problem. I found this when I was doing smartphone work in emerging markets as well.
00:06:36 Mark: You think everyone is like the people immediately around you. We're human. That's how we think. Whereas there's an entirely new set of behaviors out there from a new set of people who have not gotten access to this technology yet, and you really have to think about those people, not about yourself.
00:06:51 Mark: It's a very difficult problem to solve. It's sort of an anthropological problem if you're not building for yourself. And that is actually the stage we're at right now with these, you can call them wrappers or applications. We really have to think about the next wave of people coming on and make this very simple for those people.
00:07:06 Noah: Let's dive into a little bit specifically what you guys are working on with Knapsack. You've mentioned private AI and I'd like to touch on that, but do you want to give just a 30 second elevator pitch on what Knapsack is?
00:07:18 Mark: Sure. Knapsack is a way to use AI with your sensitive data without having to store it on an AI cloud. And this is something we bumped into ourselves. I started a company called PayJoy in 2015 with some friends. It's a fintech.
00:07:31 Mark: And in the context of that fintech and by talking to a number of other fintech founders I found, people were very cautious about actually using some of the out-of-the-box tools because it required trusting a number of cloud providers with your data in a new way.
00:07:46 Mark: You might trust them to store your data but really, with AI, you're trusting them not only to store, but also to use your data, but not basically use it for other purposes and not leak it. It's a very big leap for people, especially if they're trying to be compliant.
00:08:01 Mark: And there's really no reason to have to bundle the storage of data and use of an LLM. And so we are designing a system where you can actually run AI privately, run AI, specifically AI automations, on device or in an ephemeral cloud. That's where you upload data, but it gets deleted immediately. And we're running automations because people generally want to solve problems rather than have to write complicated prompts.
00:08:29 Mark: So we're enabling meeting prep. We're enabling meeting recording, automated meeting follow up, a whole bunch of automations that we find people in financial services, specifically, initially financial advisors who are the initial customer group we're going after.
00:08:44 Noah: Gotcha. Okay, that makes sense. And they, I assume, of the industries are one of probably the most heavily regulated as far as communication and customer contacts that go out the front door.
00:08:55 Mark: Yeah, it's really fascinating digging in here because they have access to a lot of very sensitive information. They generally have access to your Social Security number, tax returns, information that you would generally not want to upload into ChatGPT. You would not want to have them upload this into ChatGPT.
00:09:10 Mark: And on top of that, like you say, they have a lot of compliance constraints. And so there are different layers of financial advisors. So at one layer, they basically have licensing at a state and federal level that basically bars them from sharing information externally from their clients.
00:09:26 Mark: And then the second layer is RAAs often have additional constraints from the SEC around books and records where they are required to save, for example, all their meeting transcripts have to be saved at least five years from the last time they moved funds or made an update to whatever activity they're doing on your behalf.
00:09:43 Mark: And that is actually a big problem, for example, with meeting recorders, because meeting recorders don't automatically do that. If you don't pay your subscription one month, you might lose all your meeting transcripts. And that is actually a massive compliance concern for these folks. And so having some self-custodied version of AI where they have a lot more control over their data is extremely appealing.
00:10:04 Noah: I was trying to do the calculation on how much Nvidia lost because it's still down 15% from a month ago. But the big DeepSeek question. So DeepSeek, you've been in the space in talking to customers, I assume also walking them through some of the problems that maybe some of them didn't realize with uploading their own data to ChatGPT.
00:10:22 Noah: Well, first off, what is DeepSeek? Because it's not just a model. It's a company, an app, an API. Do you want to give just a quick little, what is DeepSeek? And then with all of the conversation that's been happening on DeepSeek, has that educated your customer base that you're reaching out to you? Or have your conversations changed about what you guys do versus two weeks ago?
00:10:43 Mark: I'll answer the second one first. The financial advisors have been very sharp on this in that they have not generally used ChapGPT or other things because they are very wary. They have not upsold themselves into Copilot or Gemini or these tools that are being offered to them because they are very wary of what might happen to that data and how it might actually leak out at some point in the future. There's been an average of one massive data leak every month in the last year. And so I think they're very wary, considering it's an existential threat to them.
00:11:11 Mark: But in terms of DeepSeek, it's interesting in that there is obviously the open source model that they released, and then there's the app they released as well. So the open source model is out in the world, people are using it.
00:11:25 Mark: And then there's the app, which is number one in the App Store, or has been. I don't know if it still is. And that ingests your data, and they log your keystrokes, and it's been reported the Chinese government has access to that information.
00:11:38 Mark: And Perplexity hosts a version of DeepSeek, and so you can use DeepSeek without using the app. And that's probably the best encapsulation of what we're talking about here. People want to use DeepSeek R1. They want to explore it, but they generally are concerned about using it in the app store.
00:11:56 Mark: And I find that interesting because you might be very concerned about the Chinese government getting your key strokes of whatever you're asking about, but it's probably more likely you're going to run into a problem with uploading sensitive information like a Social Security number into a US service and having a data leak from that US service than that the Chinese government is going to do something interesting with your Social Security number.
00:12:17 Mark: And so I would say people should probably have the same level of wariness with their usage of ChatGPT or Google Gemini or Claude as they have with DeepSeek when it comes to their very sensitive data. And there probably just needs to be a much better architecture where you have a lot more control of your data.
00:12:38 Noah: That makes sense. Okay. So when I think about OpenAI, or most people when they think OpenAI are really just thinking ChatGPT, this interface that they use on their phone, what is the tiers and then how does that relate to your product?
00:12:52 Noah: So you've got the LLM, the actual model that they're working on behind the scenes, which for OpenAI is not open, which is a fun irony. And then you've got this API layer, then there's an app layer, and then there's the actual interaction with the user.
00:13:05 Noah: If I have a report from a customer that I want to analyze or whatever particular example I have, and I'm interacting with OpenAI, which could be again, Perplexity or anything else, how is that different than if I'm using your app?
00:13:20 Mark: So right now with OpenAI, there are two ways people can interact. One is the way that's cheap from a financial point of view, but expensive from a privacy point of view, which is to just upload the data into the ChatGPT website or an app and use that full stack.
00:13:34 Mark: And so at that point, you're just sharing. There are no guarantees around training, really very few guarantees around what might happen to that data. And that is what a lot of people are doing right now. They're just experimenting with the out-of-the-box application.
00:13:46 Mark: The other way to do it is the way that is more expensive financially, and that is where you're using something called OpenAI Azure Service, where they set up an instance of OpenAI on a private cloud on your behalf. And this is what the Fortune 500 are buying from OpenAI, on average of a million dollars a year.
00:14:04 Mark: That's the typical cost for the service. JP Morgan and Morgan Stanley have done this for their advisors. And that allows them to have some guardrails in terms of where the data goes. And they have assurances and contractual assurance for OpenAI that the data doesn't go anywhere else.
00:14:19 Mark: What we're doing at Knapsack is taking that second scenario and making it available to everyone. You should not have to worry about where your data goes if you want to use data with the latest models. And so effectively, we're letting you take the latest models. We use open source models, sync your data up with it, and then not worry about it lingering out in the world.
00:14:41 Noah: So how is that actually working? Walk me through an actual interaction point. Is it storing the whole transcript locally and now that's on my machine? Is it just processing it live? Is it sending it all to your cloud but you're not saving it? What is the actual perimeter around my data if I'm using the product?
00:15:00 Mark: Yeah, great. Let's go through a typical user flow. So user signs up, they would sync, let's say, their Google Calendar and emails and drive into the application. That is then synced to the application. We don't store your emails or your calendar events ourselves. And this is a desktop application, to be clear.
00:15:17 Mark: And then let's say you have a meeting in your calendar. Our initial product is a free meeting transcription tool that is fully private. So you see the meeting. You start recording in the meeting. The recording happens on device.
00:15:32 Mark: The transcription happens ephemerally, which means that the recording is sent to an inference cloud. We use Groq, G-R-O-Q. And then a transcript is generated. And then both are deleted from the Groq server immediately. And then, of course, an AI summary is generated. And then that is deleted from the Groq server immediately as well and sent back to your computer.
00:15:52 Mark: We ourselves don't host any of this. And then, of course, we have other automations like meeting prep, meeting follow-up, and those are generated as well, ephemerally, where you effectively get the output, but you don't have to worry about the data being stored anywhere else.
00:16:07 Mark: And that is using your data. And your data is stored on your computer. So we actually embed and vectorize the data on your device, which means you don't have to worry about that data leaking at any point.
00:16:20 Noah: Okay. So when I think about some of the other products I've used before I got introduced to you guys last year, like a Fireflies or an Otter, you name it out there. One, I don't think I ever installed something. That was just a cloud thing that I integrated with Google.
00:16:37 Noah: Which leads me to the other fun story lately on DeepSeek. Was not just does the Chinese government have access, most likely. But also there was a completely open server sitting in the wild that some security researchers found that have all the prompts, all the users, all of the data that was getting uploaded.
00:16:54 Noah: So to your point, it's not just some state actors to be worried about with the data. It's just normal cybersecurity issues that we still seem to be having issues on a monthly basis. If you guys were to get hacked in some horrible situation, which probably will never happen, they're not getting anything of mine versus if I have one of those other tools that's cloud hosted, are they getting all of my data for the most part?
00:17:17 Mark: That's exactly right. We're going through a SOC 2 process right now. We went through HIPAA. So we've had to be very clear about what we are and aren't storing. So from our perspective, we're storing the bare essentials to run analytics, like your email address and basic diagnostic information about your general profile.
00:17:36 Mark: What we're not storing is your calendar events, your emails and for financial advisors, the important thing is we're not storing any of their client sensitive data. That's really important. And I think what you're saying about DeepSeek is really important.
00:17:49 Mark: So there are multiple layers to this. One is they themselves need to have great infosec and they don't always have great infosec. Systems are fallible. People are fallible. And we've had hacks over the last year, for AT&T got hacked and a lot of information got out. You might recall we had social security numbers leaked.
00:18:07 Noah: I think every social security number has been leaked now, isn't it? It's some crazy level.
00:18:11 Mark: Every social security number has been leaked. You probably get tons of reminders that your passwords have been leaked at some point in the past. And then we also have another really big leak was a company called Change Health.
00:18:21 Mark: This sort of flew under the radar, but they are actually really involved in a lot of healthcare data flowing between insurance companies. And therewas a ransomware attack on their services and a lot of very sensitive health care information got leaked. So this happens quite often.
00:18:36 Mark: And I think we went through the ‘80s and early ‘90s eating a lot of factory food, not thinking twice about it. And then in the late ‘90s, we started thinking, oh, I want to eat organic. We started thinking a lot more about where our food comes from and starting to be more discerning.
00:18:50 Mark: I really think over time, people are going to be a lot more discerning about where they store their data and how they store their data because of the secondary effects that are coming home to roost.
00:18:59 Mark: And really going one level deeper with DeepSeek, you have a different thing happening where then you have other people sucking in data. And there could be things going on which you would have no hope of understanding.
00:19:11 Mark: So how did DeepSeek R1 distill information? Very likely, the leading theory really is that they ran some distillation through Microsoft Azure. So they had some access to an Azure-hosted OpenAI model to the extent they distilled O1. And so even within Azure, which is what many consider to be world-class from a data security point of view, there seemed to be some interesting leakage going on there.
00:19:37 Mark: Now, what is the extent of that leakage? Was it just API calls? Was there more to it? We don't know, but the level of sophistication you would need to have to actually be able to understand that is really world-class, and people should not be expected to understand that.
00:19:53 Mark: I think just like you should not be expected to really understand the entire supply chain for your food, but you should be able to know whether it's organic or not organic and be able to make a call in the grocery store.
00:20:04 Mark: And I think that's effectively the direction we're likely to go, which is people using AI, they're going to ask, is it private? Is it not private? If it's not private, fine, I'll write poems on it. If it is private, I'm going to do work on my taxes, and I'm going to actually do enterprise work on that version of the AI.
00:20:20 Mark: And so I think we're entering into a stage where the awareness is going to go up, and people are going to start reacting to having a completely different view of where their data is and is not stored.
00:20:31 Noah: That's super interesting. After one of your acquisitions spent some time at Facebook, it reminds me of the Cambridge Analytica conversation. Even if you do trust putting your data with OpenAI and you're fine with them training on it, you probably never considered the fact that someone's gonna query that model, distill out their own model, which is based on the data that you uploaded.
00:20:52 Mark: Exactly.
00:20:53 Noah: And so the layers of trust, when you go from private to not private, all of a sudden there's so much that you have to be trusting that you don't even know to ask when your data is out there in the wild.
00:21:04 Mark: Yeah, and it's really interesting. When I went down this rabbit hole around meeting notetakers, I was really surprised that this appealed to me so much, considering there are a lot of meeting notetakers out in the world. People might be wondering right now, why is this anything novel about having automated notetaker?
00:21:17 Mark: So this came to me because I was in a meeting, and I saw two ghost notetakers waiting for everyone else to show up. And I reflected that if I said something right now, I would literally have no idea where that data would go and how it would be stored and where it might end up, because I didn't sign up for these products.
00:21:35 Mark: Then on top of that, these notetakers themselves, there are many layers of data handling that happens. There are many sub-processors. You're storing data with the notetaker themselves, whether it's Otter or Read or what have you. Then they are then passing the data to a transcription provider, so they're taking the audio and they're passing it to someone who puts it into text. And then they're taking that text and passing it on to a AI cloud of some kind to create AI summaries. Then there might be other things going on.
00:22:02 Mark: Every single one of those people has the full transcript. In some cases, the full video. Do they store it? Do they not store it? Where does it go? If I said something I regret, will it ever be deleted off the Internet? Probably not. And you don't know where it's going to show up.
00:22:17 Mark: And it seemed like an odd situation to be in where, in some cases, you don't even know what the product is being used. It's not even branded. So it seemed like a very odd situation where some really important sensitive information was out there without any sense of what's happening.
00:22:32 Mark: If you use Facebook, you at least know Facebook is the one who has this information. If it gets leaked to Cambridge Analytica, I know who to talk to, or I know who to get angry at. With the notetakers, you might not even know where this goes. And so it also seemed quite unnecessary, considering there is a private option here. And so it seemed like there was a better way.
00:22:52 Noah: It almost makes you joke, instead of this podcast, you could have just had your AI summarize your LinkedIn and my AI summarize my LinkedIn and have a back and forth. I remember in the early days when we found out not all AI was AI and most AI was actually humans overseas paid for by Silicon Valley investment dollars.
00:23:08 Mark: That's right.
00:23:09 Noah: I remember the days of the meeting chat bots, maybe six years ago. And I remember laughing because we ended up on one and it turned out, I forget what company it was. It was just people. But their chat bot started talking with our chat bot, and we both used third party systems that were just trying to negotiate with each other. And it reminded me of those Waymos that are just stuck in a parking lot and have no idea where to go.
00:23:33 Mark: Yeah. That's an interesting feature where you nominate an agent to go to a meeting on your behalf and give them a directive and see what happens.
00:23:40 Noah: Yeah. So you're sitting in the middle of a lot of this, where do you see some of this stuff going? What's your controversial take? Or what's your counter take that you're seeing from the inside of where does some of this usage go? Does it go as far as people think? Do we need to be talking UBI in the next six months?
00:23:57 Noah: Is it, no, the people that you're actually working with boots on the ground are still waiting to see what happens. Where do you see this AI, whether it's meeting bots, whether it's transcriptions, but the interaction point between your everyday consumer and AI, where do you see that going in the next 2 to 5, 10 years?
00:24:15 Mark: I think it's very easy to jump from seeing AI write a lot of great software code and basically independently maybe even do PR's review code or ship software, and jump to, oh, they're also going to take over business processes. They're also going to replace a whole bunch of other white collar work.
00:24:35 Mark: I actually think there's a little more of a climb there in terms of actually making that work. You're talking about very different personality types in terms of being willing to relinquish control.
00:24:46 Mark: Software developers have a very high propensity to want to try a new tool and try to help that tool help them be more productive. I think if you're talking to someone like a financial advisor, someone in the healthcare system, someone in the trenches in some of this white collar work, they're going to want to retain a certain amount of control. Their ass is on the line effectively.
00:25:04 Mark: And ultimately, our view is that it's going to start with automations and lead to agents when the automations can be orchestrated, and the human is going to remain in the loop. And the immediate thing we're going to see is that the humans are going to be a lot more productive.
00:25:19 Mark: The area that we've been spending a lot of time on is really in relationship management. So there's been a lot of talk about the systems of record and the fact that you could actually create using code development tools and other tools, you can actually replace the systems of record. I know Klarna has talked about doing that. So I think that's really interesting. That's possible. You can save a lot of money.
00:24:38 Mark: What I'm more interested in is the systems of intelligence that are sitting on top of that and the systems of insight. And those systems are very underdeveloped. And a great example is just optimizing your network. A lot of financial services runs on relationships and data. Effectively, what you're doing is you are reaching out to people and seeking opportunities and optimizing your opportunities through those relationships.
00:26:05 Mark: Even VCs, a lot of GPs, we talk to spend their time reaching out to other GPs, making sure they are in touch, understand deal flow, and of course, reaching out to founders, making sure they have good deal flow of founders and they're helping their existing founders as much as possible.
00:26:19 Mark: And having a system of relationship intelligence on top of that is actually a very difficult task. I know Affinity has tried to do it. Other people have tried to do it. Salesforce has tried to do it. The CRMs have generally failed. Within financial advisory world, CRMs have a worse NPS than Comcast. They are required from a compliance point of view, but they are absolutely hated.
00:26:40 Mark: I think that is interesting because these CRMs have tried to provide this level of value add, but they have not been able to do it. And I think that's where the AI automations are really going to take off in optimizing your network, optimizing your relationships, making it a lot easier to generally just be better with relationships.
00:26:57 Mark: And whether that's initiating, whether it's following up, whether it's coming to a meeting more prepared, whether it's effectively reaching out to someone because the Warriors won something last night, and you know they're a massive Warriors fan, and you want to write that text or that three sentence email just to remind them that you're in their thoughts. That's something that is going to be a lot easier in the future.
00:27:20 Mark: And I think for the non-technical business crowd, especially within finance, those kinds of automations are going to have a really big impact, and it's probably where a lot of this AI automation work is going to surface and provide the most value initially.
00:27:36 Mark: I think then there's the tedious, dull, dirty and dangerous, as often people say in terms of automations. When it comes to financial services, there's a lot of tedious form filling and other work that I think will be automated, but again, with human supervision.
00:27:51 Noah: It's interesting. One of the things that made me chuckle was when Salesforce, I mean, they announced AI every six months, but their latest AI tool for salespeople to automate the outreach process. They at the same time posted, I think it was 150 net new jobs that they were hiring for for salespeople to sell the tool that uses AI to replace salespeople. If you want your short-term bear case of we're not quite there, the people selling the tool to replace salespeople use salespeople still.
00:28:21 Mark: Exactly.
00:28:22 Noah: So what is the case that you think people don't think is actually as big of a deal as it actually is? Is it the tedious work? Where's the under-leveraged part of AI in your customer's world today?
00:28:35 Mark: So we're spending a lot of time with financial advisors. Most financial advisors are not using AI at all.
00:28:40 Noah: Oh, interesting.
00:28:41 Mark: By the way, they don't use Salesforce either. So a lot of people ask me about Salesforce, Agentforce. Why isn't that a factor here? It's not a factor because they have very limited penetration within this world. Financial advisors often like to use tools that are built for them. The 2 top CRMs they use, Wealthbox and Redtail, are ones that most people hadn't heard of.
00:28:59 Mark: They use tools like eMoney, RightCapital. These are tools that are very verticalized. So for the most part, they're not using AI very much besides maybe asking some high level questions about some funds they're looking at or things like that. And in many cases, they're not using meeting notetakers either.
00:29:16 Mark: So I would say meeting preparation, meeting recording, meeting follow up and then this relationship intelligence work of actually figuring out how to optimize their network, those are things that are underutilized right now where there's a lot of near term potential. In the long term, I think there's going to be a lot more potential around process.
00:29:35 Noah: Gotcha. Okay, interesting. We should have brought this up at the beginning, but if someone wants to go try this out, this is available publicly right now, where do they go?
00:29:44 Mark: Yeah, so our automated meeting assistant is completely free and available on knapsack.ai. So K-N-A-P-S-A-C-K.ai. You'll be able to download, use it. It's free forever, which means you don't have to worry about a free trial. And you have complete control over the saving of your transcripts.
00:30:03 Mark: And you also don't have meeting bots in your meeting, which I've heard from a lot of people is very useful. And we're rolling out automations on top of that in the near future. Right now available on Mac with Google and Microsoft accounts but available on Windows very soon.
00:30:20 Noah: Okay, awesome. All right, so as we wrap up here toward the end, thank you again for your time, couple quick questions that we like to do on this podcast just to get a sense of where you're at. Besides obviously this one, cause it's the best one, what is a book or podcast that you're currently enjoying or currently getting something out of?
00:30:37 Mark: So I spend a lot of time listening to AI podcasts, like Training Data and similar podcasts. And I realized I would get a lot more out of listening to financial advisor podcasts. So Kitces is a great financial advisor podcast that ended up being really interesting in terms of understanding a lot of the nuances of the work of a financial advisor. And he does a great job putting together that podcast.
00:31:01 Noah: Oh, interesting. Okay. If you could live anywhere in the world for one year, where would it be?
00:31:06 Mark: I would love to live in Australia. I love their attitude and I love the fact that there's a lot in Australia that can kill you. It keeps you on edge.
00:31:16 Noah: My favorite picture is there's a green Australia, like a Google Maps color green and then blue. And it says, the green is where things on the land can kill you and the blue is things in the water that can kill you.
00:31:28 Noah: What is your favorite productivity hack? And the given is Knapsack. Obviously, people need to use that. But besides your own product, what else have you got?
00:31:37 Mark: The productivity hack that I've been using consistently for the last 20 years is if something takes less than three minutes, you just do it right away and you don't think twice about it. You don't put it on your to-do list. You just avoid that completely. I was a big inbox zero person for a while. I read the book on keeping inbox zero and that was the takeaway that I keep to this day.
00:31:58 Noah: I think mine on that one is having kids. Nothing forces you to prioritize. When I was before kids, man, I'd be like, I'll just stay up a couple extra hours and knock this problem out. Now, I'm up at 6:30 whether I want to or not.
00:32:11 Noah: If people want to follow up, I think knapsack.ai is the company. They can download that. Don't need a trial. It's free forever. What if they want to find you, where can they find you?
00:32:20 Mark: They can find me @markheynen and they can find me on Instagram, Threads, Bluesky, LinkedIn, all the socials.
00:32:29 Noah: All the things. Love it. All right, Mark. Appreciate it. Thank you so much. Thank you guys for listening to Venture Everywhere podcast. And thanks again. Really appreciate the time.
00:32:37 Mark: Thank you very much.
00:32:39 Scott Hartley: Thanks for joining us and hope you enjoyed today's episode. For those of you listening, you might also be interested to learn more about Everywhere. We're a first-check pre-seed fund that does exactly that, invests everywhere. We're a community of 500 founders and operators, and we've invested in over 250 companies around the globe. Find us at our website, everywhere.vc, on LinkedIn, and through our regular founder spotlights on Substack. Be sure to subscribe, and we'll catch you on the next episode.