LAUNCH Accelerator 36 Public Demo Day

All right. Liking this easy jazz.

Oh, Will Barnes is here.

My god. How you doing, buddy?

I am.

How you doing?

Great. I'm at the ranch.

Yeah, it looks like it.

And we are live.

All right. Thank you so much, Lucas. Let's take down our music. Welcome everybody. Welcome to the public demo day for our launch accelerator cohort 36. Welcome to our founders, our guest investor judges, our syndicate members, and our live stream audience on YouTube, X, and LinkedIn. We're really excited to have everyone join us today to see these amazing eight founders show off what they're building. To introduce myself, I'm Bianca. I work on the investment team at Launch. It's been such an honor to work with these eight companies for the past few months and we can't wait to see what's in store for them. For anyone new to this, the format is that our founders are going to pitch for exactly two and a half minutes. We time them and then I'll take one question from our guest investor judges for that founder. Then at the end, we'll ask our judges, of course, to give us their top three of these eight companies. And we'll also ask our syndicate members who have joined us to vote in a poll for their top company. Before we get started, we do just want to take a moment to thank our amazing partners for the accelerator, Mercury, PayPal, Optimus Legal, and Render. We really love working with you. And to kick us off, we would love for our three guest investor judges to introduce themselves. So, we'll pass it over to William to start us off.

Thank you, Bianca. Nice to meet everybody. Some familiar faces from San Francisco. My name is William, originally from the UK, been in the US for about 20 years now and I run a fund called Carmon Ventures. Invest in pre-seed and seed kind of application layer and critical industries and defense tech.

Wonderful. Thank you so much William and Julia.

Hi everyone. So good to be here. I'm Julia Kger. I run a venture firm called Perry Pursue. We invest a bit more cross-stage, cross-sector. So, from seed to growth, from CPG to longevity to defense, and have a community of about 3,000 founders and operators and generally folks that have been smacked around in the process of building something, backing companies with us.

Wonderful. Thank you so much, Julia and Dee.

Hey, everybody. I'm a partner at Menllo Ventures. Used to be a founder in my past life of a company called Glean. I do early stage investing here at Menllo, but we do sort of all stages. Most notably, we were early in Anthropic at the Series C and led every single round since. So, really good to be here.

Awesome. Thank you so much. And Jason, any words before we get started?

All right, we're going to get started with our first founder. We have Chris at Argonoth. Can we hear you, Chris?

Yes. Testing. One, two, three.

All right, three, two, one, go.

All right. Hey, everybody. My name is Chris, co-founder and CTO of Argonite, and we use AI to solve cyber security and compliance of software delivery in the DOD. My co-founder, Bo, and I have more than 25 years supporting aerospace and defense projects at Redstone Arsenal. And that's why we know you can't vibe code for a Blackhawk. Today, it takes 3 to 7 years, millions of dollars, and thousands of man-hours to ship and certify new capabilities going to the war fighter. And we've all seen how AI is transforming the way we write code. This means the volume of software entering the DOD's accreditation backlog is growing literally week over week. And this is because technical teams have to manually build cyber packages from across fragmented systems, code repos, build tools. It's a super messy process and it does not scale. And every domain in the DOD has its own compliance framework. Things like ATTO airworthiness, etc. They all share the same root problem. They're drowning in data and manual work. So we make it easy to capture that critical data from user stories, code commits, test results, other parts of the software delivery process. Our agents embed directly into customers' CI/CD pipelines, gaining access to the tools, data, and workflows used all across DOD projects. These artifacts are cleaned, collected, and our agent maps the data to the relevant security controls in minutes, not months. We formed the company last summer, and since then, we've obtained the necessary contracts and access needed to service our early pilots. These pilots represent funding through a mix of SBIR and direct awards, and we anticipate closing more than $2 million over the next 60 to 90 days. Initial growth starts with these cyber funded awards, then scales into programs of record as we expand all across DOD systems with our cyber security agents. Four deployed engineers and small pilots unlock early adoption, but high margin licenses drive growth as we scale across federal programs. The DOD's 2027 budget is over $200 billion for RDT&E. We have a clear path to break $100 million in ARR over the next 3 to 5 years. Again, I'm Chris. We're Argonoth and we're automating the security compliance of software in the DOD. Thanks.

Way to start us off, Chris. Thank you so much. We're going to move over to our judges for a question for Chris, but I do want to remind our syndicate members that there is a Q&A box in the Zoom. So, you are able to submit any questions for a founder in the Q&A at any point. Founders, please keep an eye out there. And let's start with William for a question for Chris.

Thank you, Bianca. Thank you, Chris. Who's the buyer of the software and how do you help them get promoted or hit their annual bonus?

It's a great question. When we're selling, sales motions look like selling directly to the project offices. These are the folks who actually get money from Congress who have a mandate for certain areas of or certain capabilities in the DOD. We're also selling into some of the defense primes. They have an incentive to bring new capabilities to their customers. And so a lot of the teams, it's been a pretty easy sell because everybody has used ChatGPT, everybody's used Claude, they lost half their team because of DOGE. So there's really a transformation effort from the top down.

Thank you.

All right. Thank you so much for the question. And we're going to go over to our next founder. We have Jay at Loft. Can we hear you, Jay?

Howdy. Howdy. Hopefully so.

Yep, we hear you. And we see it. Three, two, one, go.

Perfect. Jay Cantino at Loft. Quick numbers. We've generated about $330,000 ARR with an average of 94% month-over-month growth since our November 2025 launch. And what problem are we solving? Meet Don. Don's buying his first house and he has no idea where to start. So, he turns to usual search sites and apps and he finds thousands and thousands and thousands of listings, but has no idea which ones actually fit his preference or lifestyle. Even when filtered by location, configuration, price, and type, he still has to try and sort through over 1,800 listings. And the worst part is he can't remove any with vetos. He can't prioritize any styles with likes. And there's no progression. Every time he visits a platform, he's starting new. And then Don found Loft. And Loft instantly begins to learn what Don likes by letting him swipe through real homes, not filters. And when I say what Don likes, I don't mean beds, baths, square footage. Loft sees what Don sees. We have a visual AI engine here that analyzes every photo on the platform, classifies them across our custom lock matrix and maps him for their aesthetic and lifestyle dimensions, which then allows to map Don across those dimensions with every swipe he makes, every photo he lingers on, every action he takes. Not only letting him find his perfect match in a fraction of the time, but more importantly, Don now also has a trained Loft Q fingerprint, a personalized algorithm that knows his housing preferences better than he does. And in a market landscape where a single algorithm can cause a global standoff between the world's two largest economies, we built the first one specifically trained for and developed on the largest asset class in the world, real estate. And as our users are finding their perfect home, they're also quietly training the market's most comprehensive and visually aware autovaluation model that's built on real-time behavior signals and localized to every market. A model that we can now license to institutional capital, iBuyers, lenders, insurers, and builders, cutting their noise and sharpening their bottom line. And who are we? That's my co-founder Raj and I celebrating our first customer at Loft. Prior to Loft, we bootstrapped a marketing agency nationwide focused in real estate. And we're taking the lessons we learned there to bring Loft to market. And where are we now? We've generated about $330,000 ARR through our initial launch in Colorado. Our ongoing expansion in California and Texas. And we're now preparing for our phase 1 nationwide deployment targeting 39 key metros across 19 states, bringing our coverage coast to coast in less than 10 months. We are Loft, the behavioral intelligence layer of real estate.

Awesome. Thank you so much, Jay. Let's go over to Julia for a question.

That was super interesting. How do you think about defensibility here? Because this is an area that there's a lot of folks that are doing stuff differently but have money that if you had a big shiny press release tomorrow could turn attention here. So curious how you think about moat.

Yeah. No, absolutely. So I think our moat is not the UI, the flashy like oh swiping dating for homes, that's like you said very easily replicatable, like instantly someone could have that. Our moat is more about the UX underneath. A lot of the incumbents, Zillow, realtor.com, they're all search platforms meaning that they show you what you ask for. Loft is a discovery engine. So our algorithm beneath it that actually is in the driver's seat showing that user what types of homes they like is really what our defensibility is. I like to make the comparison of TikTok and Instagram Reels, right? Similar concept, different algorithm, different result. So, I think the biggest part of our moat is our algorithm beneath, not the UI on top.

Great. Thank you for the question and nice work, Jay. Next up, we have David at Sighteline. Can we hear you?

Everyone.

Hello. All right. And we see it. Three, two, one, go.

Hey, everyone. I'm David, founder of Siteline, a growth analytics platform built for the Agentic web. The web as we know it is changing. Meet one of our real customers, Anna, who runs growth marketing at a dev tools platform called Servos. Their website and docs get over 5,000 visits per day from AI agents from Cloud Code that are evaluating, installing, and purchasing their authentication solution on behalf of their users. But her website and developer docs are still built for humans, and her current analytics platform doesn't show her any of this new essential customer traffic. Siteline helps businesses understand, market to, and convert these new customers as the growth analytics platform built for the agentic web. We provide crucial insights to businesses around AI agent and crawler traffic on their site, what they're looking for, and where they're potentially getting blocked or stuck. We then close the loop by showing when and where brands are cited by AI apps and appear in AI answers, and importantly, how this connects to increased new customer signups, installations, and revenue. From these insights, we then provide tailored recommendations for tech optimizations and content strategy, as well as done-for-you placements on high authority websites. We're a team with strong backgrounds in data science and growth at companies like Twitter and Skyscanner, and since launching three months ago, we gained over 300 signups on our free plan, have 100 websites integrated sending us 3 million agent events per day, and are driving $53K in ARR from 28 paying customers. What we're hearing from them, especially the ones with technical target audiences, is that their organic growth strategy is being completely upended. They're flying blind as human visits to their websites and docs are way down despite adoption of their product soaring. Nobody likes to read docs, so they're sending their agents who are struggling to navigate through walls of unoptimized texts written for humans. If 2026 has taught us anything, it's that the agentic coding revolution is real and here to stay. Over half of all code on GitHub is now AI generated and coding tools like Cloud Code and Lovable are already a $5 billion market growing at 38% a year. The discovery and seamless integration of building blocks by AI is becoming the new growth channel for B2B software. Yet incumbents in the space are focused on blocking or filtering bots from their metrics rather than leveraging them as the customers of the future. We're Sighteline, growth analytics for the agentic web. Thank you very much.

Awesome. Thank you so much, David. And just a reminder to our syndicate members, there is a Q&A box if you have a question for a founder at any moment. And let's go over to Dee for a question for David.

Awesome job, David. This is obviously a product that many of our portfolio companies could use. I'm curious like with the customers that you work with, what are the ROI metrics that you're actually seeing that you can drive as they use Sighteline?

Yeah, I think it starts with, it's not a direct click-through as much as it is in Google search because people are sort of ingesting and understanding. So we do show increases in visibility as a leading indicator, making sure that you're crawled without errors, but really it comes down to new customers. And so that can be through clicks, but we also have customers who are AB testing certain parts of their website, certain blog posts, and holding back others and then seeing what the incremental lift is. We focus top of funnel obviously on new leads and new customers. But then we also have lower down a focus on retention as a lot of our customers want to make sure that their existing users are able to navigate help docs with agents or set up API docs with agents. So those are some of the success metrics.

Awesome. Thanks.

Thank you so much. And our next founder, we have Andrew at Yorkby. Hey, Andrew.

Hello. Hello. Can you see my screen?

I can. All right. 3 2 1 go.

Hello. My name is Andrew. I'm the co-founder of Yorby AI and we help brands go viral on social media. To give you a background on our team, I left my investment banking job 5 years ago to become a content creator and my co-founder is an ex-Google software engineer. But more importantly, we have a combined following of 1.3 million followers and have generated over 400 million impressions across platforms. And I've also personally ran a marketing agency that's helped taken a few software brands to seven and eight figures in revenue. So, both my co-founder and I have watched the creator and marketing ecosystem grow and evolve in the past five years. So, what does Yorby do? Well, we do two things today. Today, we are a content script writer and also AI video generation tool. And what you're seeing right now is our AI video generation workflow where a user simply provides an AI avatar they want to create a video with. And our AI will create a video that looks, feels, and performs just as well as human content creators. But our end goal is not just become a workflow tool. We actually plan to become a fine-tuned model for social media marketing by building a feedback loop to track what performs and what doesn't perform on social media. Our ultimate end goal is to sell API access to other AI generation tools, brands, or marketing agencies, effectively turning us into the taste layer that tells people what to make and not just how to make it. But does it work? Well, here's an AI avatar account we use to market our own tool, Yori, that single-handedly grew to over $6,000 in monthly revenue. Or another one of our customers who has been able to hit multiple viral videos and millions of impressions ever since using Yorby. Or last but not least, here's another one of our customers we just onboarded last week in which the first two videos were already gaining strong traction. We launched Jorby a couple months ago and since then grown to over $80,000 annual recurring revenue and in terms of our business model, we charge a monthly subscription of $40 for our content script writer and $300 for our AI video generation tool. Thank you. We're your AI and we're here to help brands go viral on social media.

Thank you, Andrew. And let's go over to William.

There we go. Perfect. Thank you. Love the energy. Andrew, thank you. I don't know a ton about the creative space. If everything goes perfectly for the business, could you walk me through how big you think it could be?

Yeah. So overall, we like to see ourselves targeting the broader marketing space, which is a trillion dollar industry, but we're just starting out within the social media marketing space and specifically within software apps and social media marketing. And with the new wave of vibe coders and just everyone building their own software apps, like I believe Lovable alone has over 50 million users using their product, we believe this is a growing and large market even within the smaller TAM that we're targeting in the first couple of years with social media marketing.

Thank you.

Awesome. Thank you, Andrew. And thanks, William. Let's go over to Ilia at Super Translator.

Hey,

hello. We hear you. And we see it. Okay. 3 2 1 go.

Hi, I am Ilia, CTO of Super Translator. We translate animal to human language. We make your dog talk in human voice for real. Hi, I am Ilia, CTO of Super. Actually, I'm CEO of Super Translator and I'm an award-winning machine learning engineer. Worked at Cisco, at Cantina. Our CEO Jim was an architect of part of Windows 10 and our science officer Dr. King is a famous linguist who decoded a real language and he is a professor at Stanford. Here's how the app works. You just point your phone at your dog and you hear what the dog is saying but in human voice. This is a real working app by the way. So how do we know it works? Mostly by science. We programmed actual rules from the thousands of peer-reviewed articles on human communication and we inputted knowledge from linguistics, cognition, acoustics and anthropology. Animals don't communicate like humans. They communicate through multiple body signals. And we use AI for the world's first real translation. But the long-term vision is to use it in smart glasses and watches and cameras for an animal translation platform like OpenAI. OpenAI is for human language. Super Translator will be for animal language. So how do we go to market? First we create demand by posting on YouTube and TikTok and Instagram. Then we launch with limited access and then we convert our users into marketers with social sharing and then expand into other animals like cats. Over half of the world owns pets and that's over a trillion dollar total addressable market. We plan to monetize by monthly subscription and contextual advertisement. For contextual advertisement, we know the unique knowledge of what your pet wants and where it actually is. So providing this unique value to pet products and services will be useful for them. What happens when the novelty wears off? We become a system of record for your dog, veterinarian history, behavior history and so on, a DogLN, and people will always continue sharing videos of their dogs and cats talking. We're mainly competing against university labs. But we have best in market model of proprietary label data, working VLM and unique domain knowledge. We have the only working translation solution. Give animals a voice. Invest in making the animal world better.

Thanks so much.

All right. Thank you so much, Ilia. Let's go over to Julia.

Thank you, Ilia. I'm curious, what does an exit look like here? Who actually buys the company?

Yeah. So as it grows, I will hope for an IPO or some platform purchasing us. So yeah, as we talked, the platforms would be the key for this transition.

All right. Thank you so much. Thanks Ilia. Thanks Julia. And next up we have Jack at Grapple. Hey Jack.

Hello.

You're quick. All right. Three two one go.

Everyone, my name is Jack. I'm founder and CEO of Grapple. Grapple is analytics infrastructure for humans and agents and we're on a mission to beat Snowflake. Meet Jeffrey, senior director of FPNA at Bloomerang and Happy Customer. He has questions about his data like how many deals will we close this quarter and are we closing support tickets fast enough? He can answer these questions pretty easily with ChatGPT, Claude or maybe cobbling together a spreadsheet. And this is the world of super simple data analytics. But lurking beneath the surface are much harder data questions. Should I worry about customers who are new next quarter? And what's our revenue retention? Jeffrey has to go hire an army of technical data engineers that cost millions of dollars. They'll go buy software like Fivetran and Snowflake and still after waiting months, he doesn't have the insights he's after. Turns out data infrastructure is a big business. All these companies were founded over 10 years ago and have built a sizable business. But all in for a data engineer with software, it's about $300K per year. Operating these tools requires deep technical expertise and cobbling together the human glue between each of these layers is slow in an AI world. Grapple disrupts the traditional data stack by merging all of these complex tools into one through an agent-first all-in-one platform. And we've built the industry's first contextless architecture. So, we figured out a way to do analytics without ever sharing business data with third party LLMs. And the best part is no data engineering is required. Jeffrey can do all this himself. And here's the product in action. We call it do-it-yourself data analytics. Way better than self-service. Jeffrey can connect his data sources like Stripe and then dive in and do things like calculate net revenue retention. This goes beyond just talking to your data and gives your entire company a governed shared source of truth for metrics. So, how does it work? Grapple ships with an on-demand data warehouse that manages itself to replace Snowflake. You can use our visual builder to map across and between sources and then our skills-based notebook helps you do advanced data analytics without writing any SQL. Most data tools are sold through data engineering teams to end users. But end users are the ones with the real problems. Grapple sells directly to these end users with a problem-first selling motion that helps us move more quickly with go to market. We land with Jeffrey connecting billing platforms for $4K a month. Eventually layer in CRM and HubSpot data and then finally maybe work with the customer success team on bringing help desk and product analytics into a single place together for that entire company. Our ACV is $84K a year and there's an internal flywheel effect here driving expansion and growth. Over the last three months we've averaged 75% month-over-month growth.

Usage. I don't have V on this slide yet, but we actually 2xed growth completely in May, so we're growing really quickly. We're currently trusted by some of the industry's best like Workshop and Bloomerang. I'm building Grapple because I'm on a mission to convince the world the way we've been architecting data infrastructure is wrong. Grapple is analytics infrastructure for humans and agents.