Town vs Instinct vs GrokBot | Why the AI Assistant Market Is Not a Bubble
Summary
Town's AI assistant for email/calendar automation believes it's a top-3 priority for Google and Apple within 12 months, not a bubble—the defensible category winner will have network effects at the agent level. The founder pivoted from a failed AI tax product after recognizing that no one had built great AI operating natively in email, capturing product-market fit in weeks when agentic models became capable.
Key Takeaways
- Pivot from boring finance automation (AI tax prep) to email/calendar AI after 1 year of failed traction. The key insight: massive TAM of mainstream users running businesses through email, yet no dominant product existed. Achieved PMF in 2-3 weeks with a prototype.
- Timing matters more than customer discovery: waited for Opus agentic capabilities (Nov-Dec) rather than surveying 100 customers. Built for themselves first, leveraged OpenAI launch momentum simultaneously. Technology enablement + market excitement = faster PMF.
- Frame cannibalization fears as a TAM expansion problem, not a moat problem. With 1B potential paying users and only 100K-1M users captured, focus on market penetration speed over defensibility. Defensibility discussions are 'a luxury' for less successful companies.
- AI assistant market winner will have network effects at the agent level, not just model quality. Distinction between open-source (tinkerers only) vs. productized AI (mainstream adoption). Accessibility architecture is the competitive moat, not the model.
- Founder self-awareness on building: 'you're always embarrassed about your product at all times' and 'dark thoughts' about incumbents at 3 AM are normal. Daily active user (DAU) validation from investors/peers matters more than perfect positioning—ship imperfect products to real users early.
Related topics
Transcript Excerpt
I know what I'm building is a top three priority at Google and Apple like in the next 12 months. Not a top 10 priority, like a top three priorities. The product in this category that will win will have a network effect at the agent level. Now, the hottest category in Silicon Valley is AI assistance. On the consumer side, you've got Instinct. On the enterprise side, you've got town.com, founded by today's guest, Jean Denise, formerly CTO at Plaid. And this conversation today is probably one of the most pertinent discussions that there is. I think you'll trust your agent to decide what data to share with other people without you intervening in 5 years. You can build now at the speed of machines, but you can only learn at the speed of humans. I don't think Instinct and Town are trying to do t…