Episode Transcript
[00:00:00] Speaker A: Foreign.
Welcome to the Growth Focus podcast. I'm Gary Lafty, your host and founder of the Growth Focus Partnerships where we help MSP owners and B2B tech companies fix the revenue gaps that are quietly complex, costing them growth.
Today's guest is Mark Haynam, founder and CEO of Knapsack. And if you're still trying to figure out where AI actually fits into how you run your business, then this episode is going to cut through a lot of that noise. Mark is a five time founder with two exits and he's built Knapsack specifically for regulated industries where privacy isn't just a nice to have, it's the actual differentiator.
But what makes this conversation worth your time isn't just the compliance angle. It's how Mark talks about growth, niching down what's worked in lead gen, what flatly hasn't, and why. He thinks the founders who thrive this year are the ones who stop asking whether AI is useful and start running agents this week, as per usual, there's no fluff. Let's get into it.
Mark, it's great to have you join us on this episode of the Growth Focus podcast. Why don't you start by introducing yourself? You tell us who you are and what is it that you actually do.
[00:01:29] Speaker B: Yeah, really nice to be on.
So I'm founder and CEO of Knapsack. Knapsack is focused on helping people in regulated industries, specifically in financial services, use AI agents privately and securely. And we've been active for about two years helping people in financial services actually be able to run agents well and easily.
[00:01:53] Speaker A: Well, fantastic. And I can't wait to dive more into it in more depth as we go through this episode. But just as a dial back a little bit here there, Mark. Obviously there's no shortage of AI tools out there right now. Everyone listening just knows that we're nodding away. What was the gap that you saw that made you think this has just got to exist out there?
[00:02:11] Speaker B: Yeah, so the original concept of Knapsack came from my own experience. I'm a five time founder, started, you know, five companies with friends, two exits, two profitable companies and my most recent company is Payjoy. That's where I met my current co founder. Payjoy is a, you know, effectively a device financing company, so finance firm. It's doing quite well.
What we found when we were trying to implement AI at pageoy is that there were some data security issues. You know, we wanted to make sure that there was a lot of transparency on where the data was running, where we were able to make sure that we as a regulated entity, we wanted to actually make sure that we were attentive to all of our licensing requirements. And we found context was super important.
You know, there's a lot of nuance in how we deal with issues, especially within financial services. And so actually having that context be incorporated in whatever AI systems we implement was really important. So we did two things. We implemented a private system that was compliant. And the second thing we did is we made sure it was completely connected to all of the data sources, including data sources that are not actually generally connected to the horizontal tools, whether that's Copilot or ChatGPT or Claude. And so we did that extra work. Some of it's quite painful work to actually connect those tools.
And as a result we get a very effective system, a system that can actually run agents on a regular basis in the background to do all of the primary things that people have to do within a firm, whether that's client facing work. So meeting, prep, meeting, follow up, automatic drafting of emails, or actual internal work, you know, making sure that you're aligned to your goals. So those are generally the things we did to make sure that we solved that root issue. And, you know, it's been working very well so far.
[00:04:07] Speaker A: Yeah, well, as I said, there's so much more I want to dive into, especially when we go into things like compliance and what's making you different.
But before we do that, you made a deliberate call to go narrow. You know, go financial advisors, advisors, for example, in a sea of industries that you could have gone after, especially as you said with your experience with your exits, how did you validate that this was the right industry to go into the right niche for you?
[00:04:30] Speaker B: Yeah, it's a great question.
So financial advisors have a few properties that are very interesting. So one property which is something we noticed right away when they were approaching us, is that they are flooded with meetings. They really have this really extreme sort of situation where they have an average of 120 clients.
Each meeting is high stakes.
They are heavily regulated. So they need to make sure they do right by their clients and are diligent in their follow up. And there's heavy fragmentation of tools. They just have a lot of tools that they work, work with and they really want to have these tools be, you know, basically easier to navigate. So they're tired of the tabs. They have 15 to 50 tabs open. They don't want to navigate through these tabs. The other property financial advisors have is that they're mostly independent businesses. Some are under a Broker, dealer. So there's a different process there, but a lot of them can buy technology right off the bat. And so that actually was very useful when we were first starting out as an icp. You know, I like, with a lot of my companies, I always like to narrow down on one particular customer archetype so you can really get to know them very well. And as I'm not a financial advisor myself, I really had to go very deep. So we assembled an advisory board, went very deep in understanding what these advisors do, who they are. And what we found was, you know, the need was very strong. And in fact, we were not alone in trying to address this need. We've had. We've seen other companies, you know, raised significant amounts of money, Jump and Zox being the two big ones, basically building note takers in this space.
But for us that was actually just a sign of encouragement because that means that this is a space where there's going to be a change in behavior and we're going to be able to double down on it. Since then, we've actually broadened to other financial services firms and financial professionals. So we work with Savanto, which is a large CRM for VCs. We also work with other firms who are under NDA, who are broadly more financial services firms that really like our ability to do better at, you know, client service. So automated document generation, automated proposal generation, automated follow up, you know, enabling agents that are effective and also compliant.
[00:06:46] Speaker A: And they're talking about compliance. I mean, this is the key thing, especially in a heavy regulated industry that you're in. You've used privacy, you know, in the times we've talked before, you've used privacy as a differentiation.
And that's quite a tricky sell because it really does sound like more of a compliance checkbox rather than a growth strategy. But you've made it work. How do you make it land commercially?
[00:07:05] Speaker B: Yeah, I think the primary thing people are concerned about when it comes to compliance and privacy is data leakage. So they really don't want sensitive client data to be used to train AI models. They don't want it to show up somewhere where it shouldn't show up. And the first thing we did is we enabled a light data footprint, so minimal data retention, leave no trace. So within 30 days, by default, with our online tool, all the data is deleted. And so you can use it, but then you store it within your existing systems of record. We're not actually looking to create a new system of record. We're not looking to create a new source of Liability for anyone. The second thing we did is we created a desktop app that had fully private features where you can run AI completely locally on your computer, including a fully private note taker where no data is ever stored on the cloud at all. So this is actually leaving zero trace right away. There's no even 30 day period at that point where data is being stored.
And we found that was something that a lot of people are excited about because they don't want to be worried about their client data or other data being in some system.
Finally, what we found was people are currently thinking through which models they want to use. A lot of people rushed in and bought OpenAI's enterprise plan. Now they're rushing over and buying Claude's enterprise plan.
Google and Gemini are throwing credits out. And so now people are saying, okay, I want to actually try that one. And each of these have different environments. Now they're uploading their information to all these different environments. The information's fragmented, different things know different things about you, different tools know different things about you. And it's a complete nightmare. Nightmare. Right. And so what we wanted to do with our desktop application is give you full control over your context so you can easily switch back and forth between Gemini, chatgpt, Claude, even a local model. Right. If you wanted to use local model, you can just switch back and forth, even prompt for prompt, you can switch back and forth between those. But your context is your own. You keep your context with you no matter which model you use. And that means you are completely agnostic. You are truly model agnostic.
And that is exciting, I think, for a lot of people, because that context, I think is the most valuable thing. So the business context of your decisions, who you are, that is extremely valuable. And I think people should have a lot of control over that.
And so with our desktop app, you have full control, in fact, that is stored on your machine. Even Knapsack doesn't have access to that information.
So people have full control at that point. And we found that is the most private and therefore often the most compliant way to operate.
[00:09:52] Speaker A: Yeah, and I can see why, you know, especially as you say, in the heavily regulated industry that you're in, making sure that privacy is front and center when there is so much hype out there with all the models that are in the market. And as you quite said quite recently, people just jumping ship, they don't know which one they're looking for. They don't really know what. They're just hearing stuff on the, on the grapevine. I'm going to jump from this to now Claude, and then to the next one. But that being said, you're selling into a cottage industry, you know, and they have by, by history deep suspicion of brand new tools. They know how to do their stuff. This is how they've always been done. What mindset shift would you say, Mark, you had to install into your market before, before they even consider that what you've got is a good thing?
[00:10:40] Speaker B: Yeah, I would say traditionally financial advisors have been cautious and financial services firms have been cautious. I would say something's happening now with note takers that's changing things where people are now more interested in new tools than they traditionally have been. And that's a fairly significant shift in behavior. Now, I have no illusions that, you know, the 300,000 financial advisors in the US are not going to change wholesale. There's always going to be a subset of early adopters who are going to change and that is going to just be moving on. You know, we're going to have a.
Every year there's going to be more people shifting over. That's absolutely fine. Right. That's a good enough market to begin with. That said, I would say that now with our desktop application, which, you know, our desktop application can run unlimited agents on your computer completely tied to your data, and you don't need to worry about having an API, so we can actually use the browser to navigate to prepare you for your meetings.
I sent you a email this morning to prepare for this and I did not write a word of that email. It actually just researched your past podcast and it thought through what we could do together.
And it doesn't require you thinking through, okay, which services do I want to connect? Do I not want to connect? Do I trust Knapsack? It all happens in the browser. It basically emulates the browser on your computer. And so you can have an agent or a set of agents that know everything about you and can actually run operations on your behalf. And that is a really exciting new thing that I think goes way beyond financial advisors. That's something that I think anybody who does, who interacts with clients will want to have a team of agents working with them effectively. You have a dream team, right? You have not only a chief of staff, but you have a set of people constantly thinking through how you can serve your clients better.
That's a really amazing superpower to have. That I think goes way beyond financial advisors. And I do think now at this point with openclaw, with a lot of the agent tools that are arriving now, people are thinking through how can I do this? What, what is my plan going to be?
And ultimately, I think a lot of people are confused, I'll be honest. I think a lot of people. I've heard of people actually claiming they bought OpenClaw online and it didn't work, which is interesting because you actually don't need to pay for OpenClaw generally. So that tells me there are scammers out there who are basically doing an SEO scam where they're just, you know, basically collecting credit card numbers for selling OpenClaw.
Of course they're not selling the real thing.
It tells me that people are just, we're very early in this whole thing. People are still figuring it out. And as a result, our focus has been on just making it as safe and simple as possible for people to spin up five agents to help them week by week get their work done. And that is a much broader value proposition that I still think is very relevant for financial services. But I think it goes way beyond financial advisors.
[00:13:39] Speaker A: It certainly does, Mark. And I think this is where I just want to hang for a little bit. You know that you made quite a bold claim. You know, desktop agents doing half your work. You know, it's an extremely bold claim, but obviously you're a walking, talking proof of that. And that's why I like having leaders like you here because, you know, you've got boots on the ground.
As you said, it goes way past financial services. It can work in all types of industries. The way you're looking at it though, from your insights, what does desktop agents doing half the work actually look like in practice?
[00:14:09] Speaker B: Yeah. Let me give you an example of my typical workday and that might give you a sense for how this might work. So generally I get up and I see in my telegram, which is connected into knapsack, basically three things that I should be getting done today. My top priority items that I should be getting done today. Now that is something that I see right away in the morning and I can respond asking for my agents to start working on the first two of those. So let me give you an example. We just closed a fairly significant deal. I need to schedule a kickoff meeting for that. I need to figure out who should be in the kickoff meeting. We have a cross functional team and I need to prepare a kickoff document. And I can basically respond and say, yes, let's go ahead and basically reach out to these people to figure out if what their availability is, start work on the document, reach out to the client to find out when they're available. Right. So I can basically just do that over telegram. I can start basically generating those emails. Now I have it set up where it can, in my tone, actually send emails on my behalf. By default, it does not send emails without checking in with you first. So it cannot send emails or money or delete anything without checking in first.
So that's to begin with. Now, I typically do a lot of calls and meetings. Before each of those, I have a meeting prep come in, telling me what I should focus on and advising me if I need to do something before the meeting. And also throughout the day, I have notifications coming in when I have an urgent email or if there's something that I promised I would take care of in a previous meeting, I have notifications coming in reminding me, hey, do you want to take care of this? And if I say hit yes, it'll actually go and start work on that. Right? And often that will result in a drafted email. But a lot of work we do does not just involve writing emails, Right? For example, I was asked today whether I wanted to book a call to check in with a membership based organization I'm part of and whether I wanted to renew that membership fee. Now I said, actually yes, let's do the call and evaluate what value I've gotten. Right? And so it just started work on that. And then it told me, here's what you've gotten, not that valuable. Why don't we change it to a polite email saying no, thank you. And I said, yes, that's a great idea, let's do that. So you can see I'm actually just working through my Knapsack application rather than being like in a mire of tabs, right? Where my previous workflow was that I was actually hopping back and forth from my email tab to my CRM tab, sometimes to my messaging applications, whether it's Slack or something else. WhatsApp, sometimes looking into calendar, sometimes looking at my CRM. Often I would even be using other tools. You're in marketing, right? I have my lead gen tools. I have other tools that I'm constantly checking, right? So I have 40 tabs open. My mind is completely fragmented. I'm not even thinking of what my objective function is. Whereas with Knapsack, I've actually shared what I'm trying to get done. And it's constantly thinking through what are the things you should do today to help achieve your final objective. Right? Let's say focus.
[00:17:15] Speaker A: I can see a lot of people Mark, just thinking that's exactly how I need to live in my life. Because there'll be leaders and entrepreneurs listening and watching this going life is a nightmare. And we all know what the multi the multi tab day looks like. It's constantly swapping between, as you say, CRM this, tab that, sound marketing, et cetera.
That being said, let's talk about your growth. Let's talk about lead gen and growth and what's worked for your company.
What have you tried in terms of growth since you beginning that's really begin to work for you. You thought this is in your position. What's been a really good lead gen for you?
[00:17:52] Speaker B: Podcasting has worked. We actually run our own podcast called AI for Advisors where we interview people who are in the space. And that's been really useful. Obviously within the financial advisory space like you were mentioning is a fairly conservative space. People have to get to know you. And so how do you allow people to get to know you without having to fly across the country nonstop? A podcast is a good way to do that, I think. The second thing I've seen work is conferences. It's a fairly conference heavy industry and so we do fair number of conferences, we've sponsored conferences. So that's been helpful.
You know, email and LinkedIn, lead gen straight cold outreach has been on and off. It's been okay for prospecting and we have gotten some leads through that.
That said, I think people's relationship with their inbox is changing a little bit. And so I do feel like I've seen that diminish in terms of effectiveness over time. And so that's made me think that there's actually just a different approach here, which is building a higher public profile with a very unique thing to say, right. A unique narrative and making sure that that profile is, you know, high profile enough on X and LinkedIn and the places where our target customers are to be able to then break through the noise. And I think one of the challenges is within AI, generally within agents and even specifically within financial services, there's still a lot of noise. There's an incredible amount of noise. And so you need to have something that is really unique. Fortunately, with privacy, we have something that is unique and that is actually hard to replicate. You know, building a Rust application that sits on, that's compatible with both Mac and Windows, that promises completely private AI and private AI agents is not something most people can replicate. It's not something people can code up with Claude code in a weekend.
You know, it requires a massive amount of effort around not only the coding but also the signing and you know, making sure you resolve all data security issues. You have to make sure you're not heating up the computer. It's just there's a lot to it. And so ultimately that's really helpful for us because we have a differentiated message that other people don't have. We did find that we needed to have an online version of our tool so people could actually start without committing to a download. There are some people who are not allowed to download tools to their computers and so but they are actually able to, you know, sign in with Google or Microsoft to an online tool. And so we did create an online version of our tool so people could try it out.
That's called Studio, that's on our website.
So that, that's been helpful from a conversion point of view.
[00:20:36] Speaker A: I think it's really important what you said, mark, about. Obviously LinkedIn has changed a lot. LinkedIn has always been the place to go, but the relationship, as you said, with the inbox, has changed dramatically. There's a lot of, with AI, unfortunately, there's been a lot. It seems like every episode we talk about this, there's just the AI slop that's going into the inboxes and people are having to either ignore it or be very picky and choosy on the conversations they have. But you raised a really good point there. And it's something that obviously we've been saying for a long time now is that you could be the very best, you can have the great product, great service, but you're still competing with all the noise out there because you know, it's different. You know, what you've got is a differentiator, but the public don't. And unless you're going to get out there and create that brand profile, create a, a knowledge base for people, they keep seeing Mark come up and say, yeah, Mark's talking about this, Mark's talking about top of funnel, middle funnel. He's the expert in that field. That's the way to get above the noise. That's the way to start drowning out all the everyday Joes within the industry.
[00:21:37] Speaker B: Yeah, I agree. I mean, I think with so much AI slop out there, I think there's more premium than ever on authenticity. So it actually means having something unique to say and having it be authentic is actually more important than ever and it will carry more weight than ever and it'll actually go very far.
[00:21:57] Speaker A: Exactly. It's a long term game, it's a long term plan, you know, putting, putting yourself above everybody else now, months and years to come. So we talked about what has worked. Let's talk about what hasn't worked because obviously in every business venture there's always going to be trial and error into things we use and try to create some lead gen for us. What have you guys tried? That quite flatly was just wrong. It just didn't work for you. And why was that?
[00:22:25] Speaker B: Yeah, it's a good question. I think there's nothing where we've gotten zero value. I should say across the channels we've learned something every single time and we've always gotten something out of it. I will say we spent some time diving deep into conversion funnels and AdWords and seeing if we could actually get to a great CAC to LTV ratio just off of straight online ads.
I find that that is actually has limited ROI, especially for an enterprise sales B2B use case where someone actually often has to get on the phone before they want to talk to someone.
You know, in theory you could actually run ads to book a call. In practice, someone generally wants to get to know you before they book a call and they want to, you know, you're going to have much more qualified leads. We actually ended up having random calls booked with random people who didn't end up showing up. And so it worked. Technically there was a conversion at the end of the funnel, but it didn't work actually.
So yeah, I would say that's something I wouldn't spend too much time on before you have really strong product market fit.
Yeah, I would say that primary one,
[00:23:36] Speaker A: yeah, that's definitely a really strong one because there's more and more people using AI saying it doesn't work and it's just because there's no authenticity to it. It's just generalized output. And this is the key factor that differentiates those who are creating successful outbound campaigns and growing as opposed to those just wondering why they sound like everybody else. And AI in its very nature will sound like everybody else until you make it different to yourself. So let's look forward. What does 2026, the remainder of 2020 look like for you? You know, we're entering a new quarter. Where are you taking it and what has to be right? What green lights do you need to make sure that happens?
[00:24:13] Speaker B: Yeah, so I think there are two big areas of focus for us going forwards. So one big area of focus for us is our enterprise clients. So we basically close enterprise deals where we run set up agents that run privately in a regulated manner on people's existing infrastructure. And we have TurnKey systems and APIs that allow that to happen. Very quickly. That includes custom proposal generation, some document generation, connecting to internal systems and SaaS products, and just allowing people to accelerate their work through a very consultative process. That's been fantastic for us as a business. It's been fantastic for our clients because a lot of the off the shelf tools just don't get everything done. It gets them maybe 60% of the way there. They want to get 100% of the way there. They need someone to help them. So that's one big area of focus for us. I think the second big area of focus is our desktop app. We found the desktop app is a way people can spin up agents very quickly and a way they can get a lot of value extremely quickly. And, you know, for those who are technical in the audience, this is using the openclaw libraries on the backend, but it makes sure that they are safe and simple to use. It ensures that it does no, by default, no shell access, no ability to exfiltrate any files, no ability to do any of the scary things you might have heard about with openclaw, Basically only the good stuff, which is effectively enabling agents right away to help get your work done with humans in the loop. So we generally, by default, always have humans in the loop. So that desktop application, which is available on Windows and Mac right now on Knapsack AI, we see a lot of potential in making improvements to that. We also decided to open source it, which was a pretty big decision. The reason we decided open source is really two reasons. One is for security. We want to make sure CTOs and other people could look through the code and understand we're not, you know, grabbing any data. There's nothing untoward going on.
The second reason we decide open source is because we actually heard of a number of firms wanting to create custom versions of the desktop app for their enterprise with very custom features. I just talked to someone who wanted a enterprise dashboard where they could actually monitor what their teams were doing with the desktop app, which is fine with consent, no problem with that at all. They should just create a custom version of it, fork it and implement. Right. And they can do that very quickly using cloud code right now. So ultimately, we think that desktop app is going to be a major focus going forward.
[00:26:45] Speaker A: I'm sure it will. I'm sure it will. Well, I've got one final question for you, Mark.
Given the fact we're entering a new quarter, what is one actionable thing, one actionable advice or insight that you would share with a tech founder or an owner listening right now that they should do to differently this quarter. And why would that be?
[00:27:03] Speaker B: Well, I want to stay away from like the pablum that people always hear, which is like, get to know your customers, find product market fit, you know, spend half your time building, half your time with your customers. People have heard all that before.
What's different now? I think what's different now is you should set up five agents this week to make you more efficient and effective. You should figure out what five agents would be really useful and just set them up. Find a way to set them up. You can download knapsack and try to use a knapsack. If you want to use open cloud directly, go ahead and give it a try. If you're more technical, if you want to, you know, try to use Claude Cowork or something else, you're welcome to schedule tasks, you know, or if you're really interested in Crewai. That's another system people try to use for agents, but ultimately just start using these tools. Um, and I think you need to be ambitious when you start using them. Think of like what your dream scenario is. Which five things would you want to have done every day for you that would open up, free up three hours, four hours of your time. Imagine if you only had two hours a day to work. How would you use these tools? I think that's a good way to think about it. Um, and I guess on the flip side, maybe there are other things you should be doing outside of your existing company. Maybe there are three other companies you want to do start. And I think for the first time I'm hearing of people being portfolio founders where they're actually basically launching three or four different companies because they now have the capacity to do that. Whereas maybe a year or two ago the mantra was focus, focus, focus, don't get distracted, do this one thing. I think that makes a lot of sense if you don't have. If you have limited capacity. But now we have an abundance of capacity. And so now the question is, are you really working on the right thing or are you actually focusing on a local maxima? And you should be actually think about your global maxima is. So, yeah, that would be my advice for founders out there, which is a little bit counterintuitive.
[00:28:55] Speaker A: Do you know what? I think that's. I love counterintuitive. I think that's when you get some really great ideas, get the brain going. And I think people go. I mean, the whole mission for every episode, Mark, as you know, is for someone to walk away and go, interesting. Never thought of that. Before. That's really good. And I think you just nailed it. Just nailed it there, Mark. Well, I suppose all that stuff, if people want to get a hold of you, Mark, potentially work with you, become clients, just, or just know more about what you're doing, what's the best way they can get ahold of you?
[00:29:19] Speaker B: Sure. So you can go to Knapsack AI and I think you can basically book a meeting there. You can also reach me on LinkedIn. I'm most active on LinkedIn because of who my customers are and so I think I have a booking link on LinkedIn. You feel free to book some time there if you want to talk or just reach out to me directly and really happy to interact with your audience. I'm generally quite active on social media, so more than happy to engage.
[00:29:44] Speaker A: I'll make sure my team put those links in the show notes so they can come straight to you. So, Mark, all that leads me to say is thank you very much for your time today. We do appreciate it. Thank you for sharing your expertise and your insights and I'm sure there's, there's some good golden nuggets that people can take away from this today.
[00:30:02] Speaker B: Well, thank you for your time, Gary, and thanks for doing this.
[00:30:04] Speaker A: Not at all. It's great to have you on and thank you, ladies and gentlemen, for joining us for another episode of the Growth Focus podcast. And no doubt if you've taken anything from Mark today, make sure that you like, make sure you subscribe and share it because I'm sure that if you can use it to grow, others can use to grow and that's how we all grow together. So until the next episode, take care, be profitable. We'll see you on the next episode. Bye bye now. Mark really did share some great stuff with us today in our episode. But there's always that one thing, one nugget that I look for that actually is something worth talking about. And I found it and it stuck with me as soon as he said it.
And that is that the off the shelf tools may get you 60% of the way there. And most people just accept that gap. They assume that's just the cost of running the business.
Well, that's exactly the kind of thing we look at in the revenue leak audit. Not just your pipeline or your top of funnel activity, but what's happening after the lead comes in. Where are things falling through the cracks?
Where is that 60% becoming acceptable when it really shouldn't be?
It's a free 30 minute call, no pitch and the booking link is in the show notes. So that's a wrap for this episode. Send this one to one person who knows who needs to hear this and wants to see you on the next one.
