Ryan Okeefe

I am the head of product at Zogo.

And what is Zogo?

So Zogo is a financial education app that basically helps you learn financial skills in a gamified way. Think of Duolingo for learning and we basically gamify financial skills where you learn and earn and at the end of the day you get to learn financial skills and redeem those coins for gift cards or cash rewards.

Who are the customers for this?

Anybody. You, me, your mom, dad.

Primarily most of our audience is 18 to 26 but we range the gamut of anybody who has, who's above 13 and older.

What's the distribution? Is this software that's acquired by banks or do the people that use the software pay for it, how is that?

So we are a B to B to C. So it's a very unique model. We basically use more of a licensing model for our partner financial institutions who give access to their members or to their customers as a reward or a benefit or part of just compliance and regulatory.

That gives them the ability to kind of say, "Hey, here's a financial learning that is fun and engaging in a way that you can kind of grow your own financial understanding of different topics but more importantly they can then splice in their own kind of financial products in there and kind of cross promote market to that core audience.

I see. So you're providing inventory essentially. Yes. Okay. That makes a lot of sense. How long have you been doing that?

This is a year and a half now. I've been at Zogo running the head of product.

Digging it?

Yeah, yeah. It's fun. Startups are always fun.

It's a different world to be in at a startup that's underneath a private equity firm. Different kind of feeling and different ways that they kind of move and build. But for us, I feel like it's always great to be on the cusp of trying to solve a bigger problem, right?

Whereas if you're at an established place, you've already kind of got that product market fit and you kind of know the market pretty well. Whereas when you're at the early stages, you're still trying to figure out is this the right thing? Does this fit? Where's the market going? And for us it's ever changing. Every day is just like something new, a new problem coming up or a partner coming to us and saying, "Can you do this?" Because nobody else is doing this. And we're like, "We can try."

So that then speaks to I think the next thing. Being in product over the course of the last year and a half is a hell of a time to be in product with the advent of all these AI tools. And now with the question when somebody comes, "Can you do this?" The answer to that may have changed because the SDLC has changed. So talk to me a little bit about what it's been like to be in product over the last year and a half or so, compared to how it used to be.

Yeah. So to get back to where product is today and where it was in the past, it's really about, I feel like this core principles of what a product manager does every day is still pretty much the same as it was a year ago. The tools have changed and the speed to market has changed. But we're still doing the same kind of key things. Thinking strategically about the market. How do we solve problems? How do we scale it? How do we make revenue? How do we market these things? Right? How do we bring everybody still together to kind of commonly focus on that thing that we want to do? But you're right. The AI tools has definitely reinvented a lot of the gain for us on the product side before you used to have to spin up a lot of wheels to prototype things, to test things, to get somebody to give you feedback on a product feature that you're thinking about building. Now you can spin it up in a second through Claude and through any of the AI tools that are available now and get immediate feedback on prototypes. You used to have to go through iterations of, you know, I want to do this, talk to an engineering team, figure out what the timeline and the plan is to build something like that. Now you can literally spin up something in an hour or two, bring it to an engineering team and say, how do we get this out and test it with users tomorrow? That ability gives you the flexibility of being able to get better data and be more informed about the risks that you're taking with different features and different problems you're solving. But it's still the same core principles of product. We're still kind of doing the same strategy, the same focus. It's just the speed now has dramatically increased.

You said something very interesting, and I've been hearing this increasingly, which was you can now get prototypes in front of the users faster, which feels like it's moved where the bottleneck has changed. The bottleneck might have used to have been on the development, the coding side, the engineering side is backing out.

Getting stuff in front of users has always been the goal or the hardest part. You can make a product. I mean, you have a user to use the product. Now we're kind of in a situation where we need to optimize for making sure the users can actually give us feedback and build a loop. You have a distribution apparatus, and that's actually awesome for you.

How are you kind of optimizing for the user experience now?

It's a great point. Thinking about how do we continuously grow the product and capture feedback in these new loops that we're creating and making sure that you're not inundating people with a lot of slop or a lot of feature bloat and really trying to get there. It's really back to the same...it's crazy to think about it, but you're still focused in on what are the problems we're trying to solve for our person. What are we trying to do to help them move from point A to point B and reduce friction. For us, putting a feature out there or putting a prototype out there in front of a user, a bunch of users right now, is tantamount to almost putting an endless scroll feature. You could probably get inundated with a lot of features, so it's being very strategic about what things am I going to put in front of you, and is it something that you would even entertain? How do I incentivize you to do the thing that I want you to do? Now it's like trying to pull on strings of incentivization and motivation,

because intrinsically people are motivated by certain things, whether it's money or just the sheer feeling of I feel good or confident in doing this thing. So for us, it's like how do we incentivize the user to test this thing, give us the right feedback so that we can get enough of the data that we can make a conscious decision to move forward on this specific area.

One thing that I've heard from one particular startup is they built an incredible great startup idea, great automation flow. The one thing that he was having trouble figuring out, he was asking a bunch of other CTOs how to figure this out, is I need some AI to actually stand in place of my user to give me feedback. And my reaction is like, "That's the whole point, that's the one thing you can't optimize for."

Correct.

There's an inclination there, though. There's a desire sometimes.

There is, and I think that same inclination is there also on our side, right? It's easy for me just to say, "Oh, do I really need an engineer building this? Can I just have an AI tool build this thing and skip the engineer part?" But no, there is this inherent sense that engineering, I feel, is morphing as well as getting the feedback is important. If I just lean too heavily into the tool has to do everything and the tool has to give me all the feedback, then where is the human element in all of this?

We kind of miss out and now we're building features for agents to give us feedback to another agent, and we're missing out on the human element of, "Does a person even care about these things? Are they going to use it at all?" Even if an AI tool tells us the feedback would be 100% positive around this feature and they would love it dramatically. I ship it out tomorrow and nobody uses it. What good does that do me? For me, it's better to have tangible evidence and data to collect and say, "I had 10 real users give me feedback here." I can run an AI tool over it and say, "Find me the vulnerabilities, find me the security risks, find me the areas where we can improve this and scale it." But I'm not going to rely upon it to give me the actual feedback of whether it's a good feature or not because that piece comes only from asking actual real people.

You shifted gears a little bit.

With all of these AI tools and this whole revelation, what about all of this has got you excited for the future?

There are parts of it that are extremely exciting. For us, the automation piece, I think from my perspective, when I look at AI and the evolution of it, it is another industrial revolution kind of a moment that we're having. I think it's going to take us a while to figure out where it comes into play as far as the impact it's going to have on our society and on our technology and on us as a group. Some people are saying that it's going to be we're moving so fast and furious that we're going to be headed for a challenger moment where we're constantly using something but we're not knowing whether we're using it for good or bad because we're just plugging in things and playing around with it. But for me, I think that there are great parts of AI that are coming through the automation piece that is accelerating development and accelerating a lot of the code that we're doing today in a way that we've never seen before that helps us get to market faster, get things out there quicker to test it, understand whether these things are even good or bad but also try to find the risks and the security issues that we're having with some of the things that we're building. So I find that if you look at the evolution of engineering from manufacturing and the software we kind of follow the same boats of how we build things and how we scale things and manufacturing has been automating things like this for years. There's already dark manufacturing

buildings all around the U.S. where there's hardly anybody in there and it's just machines running without lights on. Are we going to be like that in the next couple years with engineering and software where you're running a full-on whole company with dark with no code being touched by any engineer, no person being in there telling it what to do and it's just doing it possibly. That's probably where we're headed. I see that that's probably an area of potential

discoverability and growth but I don't know if that's going to happen at a pace and at a reckoning that I'll probably force some of us to think about what does this mean now for us as an industry, what does this mean for how we capture feedback and whether this is a positive thing or a negative thing or how do we turn this into something that's going to be beneficial for all of us.

Well to that point, is this going to be a positive thing or a negative thing sociologically? Can the culture, economy, civilization absorb all of this stuff and with that in mind is there anything about this that's causing anxiety or giving you up at night right now?

Yeah for me the biggest piece that keeps me up is that the people who are going to benefit the most from the AI are people in Gen Z and younger, Gen Alpha.

Us at the, start dating myself a little bit, at the millennial age right, I'm like at the point where I've seen the evolution of computers. I've kind of come up with computer programming and now I'm seeing this evolution in AI and I've had my phone now come into play with us but for them I think the biggest fear for me is that they become so reliant on these pieces, these tools right, that they're not doing the critical thinking that they need to be doing every day and the more human interactions that we would have every day and how we learn. That's a big part of why I love what I do today with Zogo is it's built around education and for me curiosity and learning is like part of our everyday human ethos and for kids today, like I have two daughters they're very hyper focused on AI and if you think about where they're at in their evolution of AI and where it comes into play with them every day.

Search, school, homework, how do I get this answer to this question, how do I write this paper, right, and it's all being used to shortcut to get to like a faster solution or a faster way but if I ask them to explain how they got to that place, like how did you solve that problem, how did you write that paper, what is the core thesis behind what you were writing, that's the piece that we're going to probably be missing and that's where it keeps me up at night and I'm like how do we get back to some of this core principles of like critical thinking skills and keeping that as like an ever present thing in our education

where like to be honest a lot of kids today are just defaulting now to AI as being the thing that they learn from and I want to try to get back into like how do we capture more of that active learning and how do we bridge the gap for people to like keep using those key critical thinking skills to kind of help them like be curious and continuously learn.

I need to teach my son how to change a tire.

That was a punishment when I was in trouble and they'd be like go change all your tires. Oh wow.

I got a 14 year old, I need to make him do some stuff. I'm not sure what those things are.

What tools do you use on a daily basis?

So for me I use Claude a lot as far as an AI tool is concerned but outside of Claude we have Notion, I use that a lot for writing, I use, let's see, we have Monday for our project management task board stuff, Slack for communication

and then outside of that we spend a lot of time on trying to think of what other tools we have inside. I mean Claude code and a couple of other CMS tools that we have for building content, HiGraph and a few others but for me like a majority of my day is spent either working with the team and communicating and handling meetings. Oh, another tool I use that I really love. I don't know how many people have heard of granola. Have you heard of granola AI? This company is kind of revolutionizing the way in which you kind of transcribe and get meeting notes and get meetings more efficient and better. I found them out about six, eight months ago and I'm telling you that tool in itself has become a game changer for me in my everyday work because what it does is instead of jumping on a Zoom call or Slack call or whatever, it literally starts recording your meetings once you start talking and it starts capturing the notes in the background and then it summarizes the notes at the end and then you can override it if you want with your own manual notes and say hey, I did say this or I didn't say that or summarize these key points in an email and send it out to me and the three people that were in that meeting and it automatically does it for you. It's really been an effective tool for us. So those are the ways AI has kind of helped me kind of build out gaps. Clod has also helped us a lot. I use it pretty much strategically every day with my own individual Clod agents that I built myself for product strategy, product data,

user experience, prototyping, and also just like basic project management within my team. Like oh, what things are we working on? How do I prioritize these things and where do we see the best opportunities for us in the next like three, six months? So that's where I kind of spend most of my time in working with different tools.

How big is your team?

I have six people today.

Are you a remote or a

hybrid? We're a hybrid team where three days in the office, two days out, so engineering team, most of the folks reside here in Austin. We have one remote employee that is engineer that is south Texas and then we have a product manager who's out in Florida.

You guys grow in the States.

We're growing.

We're growing. I'll say this, like the pieces that like where we've started to like use AI now as a team has helped us kind of grow and more importantly figure out okay, here's the things that we do well. Here's how AI can come in and help us kind of like speed ramp some of these other areas, but also off-road some of these other pieces that we don't like doing on a month-to-month basis like security analysis, code analysis, like these things and QA stuff. So for us, we've been growing our team. We just added another engineer. We're going to probably add another one by the end of this year. So we'll probably round out with maybe four to five engineers on my team.

So that's actually the question I was kind of going with next is like you're seeing all these efficiencies you're up in your productivity, but you're actually adding teammates to that as well.

We're adding people because we're growing as a company. So as a company grows, we grow as well because we got to keep supporting the initial growth of where we're heading in the market and we're kind of going more upstream into enterprise markets. That kind of demand comes with more customizations, more feature-rich things that you can definitely use some AI to help you get there, but you really still need the core engineering team. And what I'll say this is like engineering is moving more towards agentic engineering and not vibe code engineering. There's two different pieces that I think people miss out on. They think that everybody who uses AI in engineering is vibe coding and that's not true.

When I think of engineering using AI, it's four things where they're submitting code, reviewing code, helping them kind of like QA stuff. And it's not just them giving a prompt to an AI tool and letting it build it and not even check the code. So there's still some manual checks and some other pieces that we do with our engineering team, with our AI tools, which helps us kind of like just move a little bit faster through that.

I think about things like Scrum is a 30-year-old piece of the SGLC. No one's built the ubiquitous AI agentic coding framework yet. Everyone's inventing it themselves.

I have been a real big fan with small startups to be lean and agile. We'll call it lean agile. But it's really like what I'll call the old Toyota Flow model of Kanban and we pull in the most highest impact valued items off the top of a swim lane and move it in and you just continuously flow that through from start to finish. And the goal for us is break things down to the smallest piece, move them as quickly as you can through the pipeline and then hopefully we generate enough flow in that evolution of us pulling something in, seeing it through really quickly, that you constantly have rapid iteration and constantly feel like, oh, we're achieving something because we're moving quickly and we're moving fast.

Last question. Claude Beto, are they not releasing it because it's too dangerous or is that all just PR?

I've had this constant conversation with my team twice so I am on the feeling of it being a marketing ploy a little bit right now. I mean, if you release it to the three highest profitable companies that are using it, I'm kind of feeling like you're kind of skewing it to one. Now, do I think that they have specific research models that are probably reserved for certain entities? Yes. Because they feel like if they were out in the wild, they could probably produce some critical issues across the board and the internet. But I don't think mythos is the key thing that we think it is and we won't know until it actually gets released out to the public. But I think what they're hyping it as is part marketing and part promotional.