Kaz Tahara-Edmonds is the CEO of Kazuo Corporation, an AI-native company building what he calls an autonomous distribution engine. The system identifies consumer product opportunities inside specific TikTok and Instagram Reels niches, ships apps in under a week using AI tooling, and orchestrates a network of more than 300 creators to distribute them. To date, Kazuo has launched six apps and the portfolio is generating over $60,000 in monthly revenue.
His path to that operation is unusual. Kaz started coding at eleven to build Minecraft mods, and by his early twenties had bootstrapped a string of newsletters, job boards, and consumer apps reaching hundreds of thousands of users, including Intern Insider and a holding company of products under KJC Technologies. He joined Entrepreneur First as a Founder in Residence before Kazuo became his full-time company, and the bootstrapped operator instincts have carried into how he allocates venture capital today.
What makes Kaz’s perspective worth hearing now is the inversion at the center of his model. Most companies bolt distribution onto a finished product. Kazuo starts with the distribution channel, then builds products designed to live natively inside short-form content. The result is a working thesis on where the consumer app economy is heading as AI collapses the cost of building software, why horizontal apps are likely to fragment into niche-specific variants, and what separates AI-native products that find durable revenue from the ones that don’t.
You started coding at eleven building Minecraft mods, and by your twenties you’d launched newsletters, job boards, and apps used by hundreds of thousands of people. What did that bootstrap era teach you that’s now shaping how you build at Kazuo?
Bootstrapping forces you to come up with creative solutions that can’t be solved with brute force spending and iterative testing. When I launched my first game at 15, I had a zero dollar marketing budget, so I had to invent the most elaborate schemes to get it in front of people without paying. It hardwired my brain to think scrappy first, and this led to over 150,000 users from no budget. That same instinct helps me find edges in more scalable marketing channels like influencer marketing. Building new systems and channels, instead of running industry standard playbooks. For example, seeding comments on TikTok videos to drive conversions and engagement, whereas most people focus on just the content of the videos. This instinct has been instrumental in scaling from 0 to over 300,000 users profitably.
Bootstrapping also taught me how much the small optimizations matter. When I co-founded Intern Insider as a broke university student, I only had profits to reinvest. I needed to ensure that every dollar we spent maximized the amount of revenue, but also focus on having a short payback period so that I could scale faster. It made me value every increase in conversion rate, every extra month a user stays subscribed, and intensely calculate risk of allocating spend to different channels. A single mistake could leave us negative for that month, and the inability to grow further. One month, I had to borrow money from a friend to buy groceries, because I wanted to reinvest all capital into growth. That month we gained over 30,000 new users.
Now that I have capital to work with, I still retain the same discipline. Aggressive scaling has become calculated, not reckless. Each optimization we make is amplified by our new leverage, a couple percent increase in conversion rate across hundreds of thousands of users is what has allowed us to grow 100K ARR month over month.
Kazuo operates what you describe as an autonomous distribution engine, finding product ideas, building them with AI in days, and distributing through short-form content. Walk us through what that pipeline actually looks like end to end.
Our system starts by analyzing thousands of influencers across TikTok niches. We look for niches that have emotionally resonating viral formats, easy to recreate formats, and low competition with other brands. Once we’ve identified a promising niche, we build viral features that can integrate naturally into the content style of that niche. With AI tools, we can push an app from idea to MVP in less than a week. Once our app is live, we autonomously distribute content formats that integrate our app to our network of over 300 influencers. Our system learns which formats, influencers, and features are working. Each iteration, our acquisition cost goes down, and our reach increases. Each week, the system doubles down on formats that go viral, influencers that are resonating, and cuts underperformers.
You’ve launched six apps and scaled them to $460K in annual recurring revenue. Which of those launches taught you the most, and what changed in your approach as a result?
Tethered, a long-distance relationships app to keep couples closer, was the most instrumental. At the beginning, we were focused on finding undervalued creators and finding cheap attention. We had the presumption that we just needed to find influencers cheaply, and they would handle the rest. What we found with Tethered is that creators didn’t have the bird’s-eye view to see what’s happening in the niche as a whole. Creators were working to figure it out individually, without learning from each other. We found the best approach is to manufacture the content strategy from looking at the performance of all creators and distribute the winning formats and styles to each creator. There are two reasons for this. The first is maximizing performance. By analyzing thousands of videos across the niche, we were able to extract formats that work much better than what an influencer would come up with naturally. The other is control, shifts and trends happen all of the time, and we need to pivot the network in days, not wait for each influencer to catch on. This insight created our most valuable edge: our automated content distribution. Each week, our system is able to autonomously find new formats that are going viral now and instantly distribute to our network of creators. After we built this, we now employ this engine with all new apps we launch.
Kazuo uses short-form content on TikTok and Instagram Reels as its primary acquisition channel rather than a marketing layer added on top of the product. Why has that approach worked, and where does it break down?
Most people think about product and marketing as separate. We take a different approach by making features of the product part of our content strategy. We create features that integrate smoothly into short form content and make the video more engaging, rather than turning it into a sales pitch about a product. Our most viral feature for our long distance relationships app is a real-time multiplayer drawing canvas couples use together. This canvas can be integrated smoothly into content, and create hilarious interactions that become part of the content. This works because the app becomes watchable content, while simultaneously showcasing the value proposition. On TikTok or Instagram Reels, no one wants to watch an abrupt insertion or blatant advertisement. This feature-distribution model has allowed us to get over 600 million views across TikTok and Instagram Reels.
The difficult part of this is that it doesn’t scale as smoothly as traditional paid advertising like TikTok or Meta ads. Each increase in volume requires more content formats, finding more creators, and managing more, whereas the primary lever with paid ads is spend and number of ad creatives. We solve this by automating the process with AI. Right now, our system is able to manage over 300 creators with little human intervention.
There’s an ongoing debate about whether paid and organic short-form growth are converging or diverging. From where you sit, watching this play out across six apps, what’s actually happening?
Short form content and paid ads are converging quickly. The best performing organic content generally works well for paid ads. From our 6 apps and broader analysis of the ecosystem, users are starting to value authenticity and entertaining content over polished value propositions in paid advertising, the same qualities that work for organic content. We’ve seen that creator-style ads with strong hooks, simple storytelling, and natural integration of products become the best performing ads for Gen Z. From this insight, we’ve been repurposing organic content that fulfills certain watch time and engagement metrics into paid ads. Organic content has become a great system for finding new winning ad creatives.
Where they still diverge is which metrics each channel needs to optimize for. For paid advertising click through, conversion rates, and lifetime value of a customer are the most important metrics. For organic, there’s a new dimension of virality. A video that gets 1 million views, with a low conversion rate, will beat a video with a 10x conversion rate that only gets 10,000 views. For organic virality is the most important metric, which isn’t a consideration for paid advertising.
You spent time as a Founder in Residence at Entrepreneur First before Kazuo became your full-time company. How did moving from bootstrapped operator to venture-backed founder change the way you make product decisions?
Since I’ve been venture backed, I’ve been focusing on building a system to ship products, rather than ship a single product. Every product I launched was sequential, using each one to fund the next. Now we started with the launch and distribution infrastructure first to create a foundation. From this we’re able to launch multiple products in parallel, and each release feeds into our system to improve our other products. This is only possible with VC capital, because before I didn’t have the time and money to invest into building the system. Each implementation had to feed directly into immediate revenue to survive, and sometimes sacrifice long term success.
Consumer AI products are being built and shipped faster than at any point in app store history. What separates the ones that find durable revenue from the ones that don’t?
Knowing what to build and how to distribute it are now the key factors, rather than technical expertise. These factors have always been instrumental, but now they are the whole game.
Software is becoming trivial to build, and with AI essentially anyone can create a functional app. This has opened up the playing field, where technical skills are no longer a moat. With every competitor having the ability to create equivalent product features, taste and distribution is the only way to gain an edge over the rest of the competition. The products that find durable revenue will need scalable systems of growth to acquire users efficiently, and they’ll need to use that growth to build a brand that resonates within a specific niche. If an app can accomplish these two things, they will be able to build a funnel that compounds over time and drives down acquisition cost to gain an edge over the competition. That’s what durable revenue looks like in this new market, not a technical edge.
Looking three to five years out, how do you think the app store and the broader consumer app economy get reshaped by AI-native production?
I believe that AI-native production will lead to the death of horizontal apps. Each app will split into more niche markets. Opposed to having one universal meditation app, there will be more variation: for example, a meditation app for athletes, a meditation app for entrepreneurs, etc. Each app will be more personalized for a specific user, which will increase retention and appeal, as opposed to a horizontal version of it. Today the broader consumer app economy is dominated by horizontal apps because building was formerly expensive. It would be economically impossible to build perfect variants of different apps for different niches; however, this is no longer the case. You don’t need millions of users with low average revenue on one app. Instead, you can create niche apps that serve smaller communities that are extremely personalized and sticky. Distribution will shift alongside this, where the winning strategy is finding niches with scalable distribution systems, then taking over presence in that niche before the competition.
For founders trying to build in consumer AI right now, where do you disagree most with the prevailing advice?
The piece of advice I disagree with the most is that you should focus on product and value proposition first, then figure out distribution after. That may have been correct in a world where building was tedious, and there was no social media to reach millions of users overnight. With modern social media algorithms, distribution and product have become inherently linked.
I’ve watched dozens of builders spend months designing a product and expect thousands of users to flock right after launch. Generally, great products take hundreds of iterations and rounds of user feedback before reaching a word-of-mouth factor or exceptional stickiness. Low-volume product iteration can be extremely misleading, as it’s not guaranteed that the small subset of users represents the greater population or the scalable distribution channel that you will eventually find. At Kazuo, we start with distribution first and then find products that align with that distribution. Distribution isn’t a layer that should be added on top of the product; it should be the starting point.
