What if every metric your product team is optimizing for is already a debt? Daily active users, monthly active users, weekly logins. They were all built for a world where humans were the primary users of software. That world is ending. In this episode of Follow the Gradient, Christian Woese and Melanie Gabriel sit down with Tony Beltramelli, co-founder of Uizard and now Head of Product (AI) at Miro. Tony has been building AI products since 2017, when his pix2code paper went viral and convinced him to turn a weekend research project into a company. Six years and 55 employees later, Uizard had over three million users and $3.5M ARR, and was acquired by Miro in June 2024. This is not a conversation about how to "do AI". It is a careful walk through the team-building, hiring, and integration decisions that determine whether an AI-native company actually ships and survives. Tony has lived both sides of the problem: building from a four-nationality founding team in Europe, and now turning a public-company platform into an AI-first product. We talk about: Why founders fall into the Henry Ford trap of assuming they understand the customer better than the customer does, and how Tony forced his team out of it The hire-only-when-it-hurts decision rule, and why it gets sharper, not weaker, now that AI agents can absorb the first wave of work How a four-nationality founding team turned Europe's fragmented talent map into an advantage by hiring per-city for what each city is actually good at What actually happens inside an M&A process: setting deadlines, leveraging investors for warm intros to executives, and telling employees last The counter-intuitive playbook for going AI-first inside a legacy company: separate the AI team, ship something end to end, then re-inject across the org Why daily and monthly active users are debt metrics in a world where AI agents are becoming the primary users of software The deeper thread here is not the technology. It is decision-making under uncertainty: when to listen to customers, when to keep building, when to hire, when to sell, and when to stop predicting altogether. Tony's framing is that the cost of turning ideas into reality is collapsing, which means the new differentiator is the curiosity to generate ideas worth executing on in the first place. Our biggest takeaways, including Tony's view on the one trait every product team needs to hire for in the next two years: https://www.followthegradient.io/p/tony-beltramelli-podcast — Where to find Tony Beltramelli: LinkedIn: https://www.linkedin.com/in/tony-beltramelli-513b1219/ Homepage: https://tonybeltramelli.com/ Miro: https://miro.com — 🎙 Follow the Gradient: conversations about building a business from Europe while staying sane. Follow us: Melanie: https://www.linkedin.com/in/melaniexgabriel/ Christian: https://www.linkedin.com/in/christian-woese/ Subscribe to our channels: Newsletter: https://www.followthegradient.io YouTube: https://www.youtube.com/@followthegradient LinkedIn: https://www.linkedin.com/company/followthegradient/ X: https://x.com/followgradient Instagram: https://www.instagram.com/followthegradient/ — 00:00 Introduction 01:31 The viral white paper that accidentally started a company 06:25 The Henry Ford trap and the cost of not listening to customers 09:07 Hire only when it hurts: the rule that scales into the AI era 11:34 Remote-first by necessity, with founders from four countries 16:50 Viral loops, waitlists, and growing to three million users on no marketing budget 19:27 The Miro inbound and running a structured M&A process 26:46 Why you tell your team about an acquisition last, not first 32:47 Going AI-first inside a legacy company: separate, then merge 37:43 Your users aren't human anymore: the end of DAU and MAU
