Zach Rattner forced himself through the door of a moving company for the 18th time, phone in hand, looking for an excuse not to walk in. Building AI products as an introverted engineer meant every cold call felt like torture - and half the people he pitched said his AI startup would never work.
Learn how building AI products with LOI validation de-risked a technically ambitious vertical AI startup, why live trade show demos using TJ Maxx furniture overcame computer vision SaaS skepticism, and how honest expectations kept AI startup customers engaged.
š Key Lessons
- š¤ Validate building AI products with letters of intent before code: Zach secured non-binding LOIs with agreed pricing from moving companies, proving demand for the AI startup before writing any code.
- š§ Do founder-led sales to keep the feedback loop direct when building AI products: Zach handled all sales until $1M ARR because he didn't want objections filtered through a salesperson.
- šÆ Target early adopters who benefit from your AI product today, not tomorrow: Yembo found customers with pain points solved even when the computer vision SaaS could only detect 10 items.
- ā” Use live demos to overcome skepticism about building AI products: At trade shows, Zach bought random furniture from TJ Maxx and let prospects try the vertical AI live, closing deals on the spot.
- š Set honest expectations so imperfect AI products don't destroy trust: Yembo positioned as a time-saver requiring human review, keeping AI startup customers engaged while accuracy improved.
Chapters
- Introduction
- Zach's favorite quote - Radiohead and optimism
- What Yembo does - AI-powered virtual home surveys
- Team composition and company size
- Origin story - when computers beat humans at image recognition
- Validating with BBB complaints and cold calling
- Cold calling as an introvert - overcoming rejection
- Customer reactions - polarizing responses to the AI startup pitch
- Using letters of intent to validate before building AI products
- First version - bumpy launch with limited AI accuracy
- Funding the business - seed round and staying close to revenue
- Managing early customers and setting expectations
- The surfboard story - when AI tagged a wife as furniture
- Managing AI expectations in sales and onboarding
- Founder-led sales to $1M ARR
- Trade shows and the TJ Maxx booth demo for building AI products
- Lightning round
- Wrap up
Resources
