We're backed by a16z speedrun to build the sales layer for social media. Here's what we're building and why.

what's changing

There's a culture shift happening right now. Businesses today no longer start by creating a website or having an email address; they start with an Instagram account. It is now completely legitimate to run a business end-to-end on social media, from the first point of discovery, capturing your leads, and then selling right inside the DMs.

Over the past year, we've seen social commerce grow at an unprecedented rate, but the infrastructure hasn't followed.

We're the first to build sales tools for social media. We're starting with Instagram, researching every profile behind every interaction with your account, from your likes, your comments, or DMs. We surface your hot leads and agentically engage with them.

These tools already exist for an email-based world, but they don't exist for the algorithm one we're shifting to.

why we're building this

Hotbox exists because the experience of getting enterprise dealflow in my DMs was a painful one for me personally. I've grown three startups from zero to millions in ARR by building viral growth playbooks from scratch. At my last startup, Respell (now acquired by Salesforce), I created a viral content series building AI agents for everyday tasks, like getting AI to call United Airlines to find my missing luggage or background-checking my Hinge date. Every video blew up and my DMs exploded, but not just with buyers. Most messages were a single fire emoji, some were questions I've answered hundreds of times, and just a few were serious leads worth 6 figures each, now buried in the noise with no way to spot them. I knew I was leaking money but had no way of knowing where or how to fix it. That's when I realized there was a massive gap.

the team behind Hotbox

I've been building at the intersection of AI and social media since the beginning of my career.

My first venture was funded by MILA (Montreal Institute of Learning Algorithms) to discover subconscious biases in the way that we engage our social media newsfeeds. It was used by PhD researchers at both MIT and MILA.

I later joined LinkedIn as a machine learning engineer to work on their platform ranking algorithm. Soon after, I had my machine learning research published at NeurIPS for building the algorithm that fine-tunes the particle beam for Canada's National Particle Accelerator.

After becoming deeply technical, I became obsessed with solving distribution for the most innovative technical products out there. I've grown three companies from 0 to millions in ARR.

I built a team full of business-savvy creators and creator-savvy engineers, and we're just getting started with Hotbox :)

whats next

We're really excited to be part of the SR006 batch. If you're coming to the A16Z Speedrun Demo Day on April 14th, come find me.

Or, if you're interested in having your mind blown, join our waitlist at hotbox.app :)