How David Shim Built Read.ai Into the Fastest-Growing Meeting Tool in the World

What founders can learn from building viral AI products in ultra-competitive spaces

In the era of AI-powered everything, one founder stands out for not only surviving the noisy world of productivity tools but thriving in it.

David Shim, the founder of Read.ai, built a product so sticky, it’s adding 50,000 new users every single day. Shim’s journey, which includes a $10M seed round and a prior acquisition by Snapchat, is a masterclass in how to build viral SaaS in one of the most competitive categories: AI meeting assistants.

In an interview with Andrew Warner on The Mixergy Podcast, Shim pulled back the curtain on how Read.ai became the fastest-growing meeting tool in the world — and why its success might have more to do with human psychology than machine learning.

“We are the fastest growing meeting note taker in the world today, and we have been for the last two years,” Shim said.

Let’s break down how he did it — and what early-stage founders can learn.

1. Don’t Build an AI Feature. Build an AI System of Record.

Shim didn’t stop at transcription or summaries — what most note-taking tools offer. He leaned into real-time sentiment and engagement analytics, turning meeting content into a dynamic productivity system.

“Imagine if you read a book and only read the quotes,” Shim explained. “You get a general understanding. But the narration layer — that’s what we add. We help you know what was interesting, not just what was said”.

This “narration layer” uses visual cues like head orientation and hand gestures — not facial recognition — to interpret attention and sentiment. It’s not just about what happened in the meeting, but how people felt about it.

That emotional metadata? It's what turns basic notes into strategic insights.

2. Ship Before You’re Ready and Way Before It’s Perfect

Read.ai didn’t launch with a full-fledged platform or feature-complete roadmap. They launched as soon as they proved their foundational AI models could work.

“If we can’t get traction with a free product,” Shim said, “how are we going to get traction with a paid product?”

They focused on user value before monetization — a luxury made possible by their $10M seed round led by Madrona.

But even for bootstrapped founders, the lesson applies: launch early, collect real-world data, and let your users show you what matters.

3. Your Product Is the Marketing, Especially in AI

With Read.ai, virality is baked into the core experience. When one person uses Read in a meeting, everyone else sees it. Some even get emailed the notes. Curiosity does the rest.

“We don’t have to spend anything on marketing,” Shim said. “Platforms like Microsoft Copilot and Zoom are marketing the benefits of AI for us. People try theirs, then look for something better — and they find us”.

That viral loop helped Read.ai break into unexpected markets — like universities in Colombia, where 1–2% of the entire country’s population uses the app.

4. AI Doesn’t Have to Be Niche… Yet

While many believe AI SaaS will fragment into hyper-specific tools for sales, HR, or customer success, Shim is betting on one intelligent layer across all workflows.

“If I only took your meetings, I don’t have enough context,” he explained. “I need your email, your Slack, your Google Drive, your Notion — to deliver meaningful value”.

Rather than siloing by use case, Read.ai adapts dynamically: if it’s a sales call, it adjusts the metrics; if it’s a team meeting, it focuses on engagement and follow-ups.

5. Retention is the Real Moat

Yes, the product is free — but it’s not a leaky bucket. Shim points to 80% 30-day retention after first use. That’s not normal for productivity tools, especially ones where AI is doing the heavy lifting invisibly.

“You can’t fake retention,” he said. “That tells us people don’t just like Read — they build their workflows around it”.

Bonus: Your Use Case May Not Be the One That Wins

The original idea for Read.ai was to help salespeople improve real-time performance. But one of their most loyal users? A person with early-onset dementia, using Read to remember conversations with family.

That unexpected use case opened Shim’s eyes to the broader value of intelligent memory systems — and inspired their direction toward becoming a “system of record for productivity.”

Takeaways for Founders Building in AI

• Launch early — perfection is not required. Value is.
• Build in public — usage feedback drives better models than market research.
• Think in systems, not features — the real moat is full-context intelligence.
• Design for viral surfaces — let your product spread itself.
• Retention is your north star — nothing else matters if users don’t come back.

Title Options