New episode with Bishesh Khadka, founder of Imprezia, a Y Combinator-backed company building Google AdSense for AI apps. Bishesh went from MIT to Amazon Ads in 2019, back when he’d never even heard of Amazon Ads, and built foundational pricing systems that manage billions of dollars of ad spend. Then he worked on GenAI ads at Meta before leaving about a year ago to start his own thing. Imprezia puts ads directly underneath the chatbot, ChatGPT-ads style, for every AI app that isn’t ChatGPT, and splits the money with them on revenue share, with no retainer. His case: almost every product that reaches a billion users runs on ads, the walled gardens own 80 to 85% of the market, and even 20% of a $1T+ market in 2030 is a huge number.
We talk about whether every model upgrade makes Imprezia stronger or weaker, why he does every job himself before hiring (you can’t lead a 2,400 Elo team at 800 Elo), solving cold start in a two-sided marketplace, why cold email doesn’t work anymore, why search is dying and where ad budgets go next, whether Google ends up owning AI advertising, how a real-time auction actually prices an ad and sets CPMs, why context makes ads worth more, hiring hunters instead of farmers, and whether AI kills the big ad agencies.
Watch on YouTube:
. More from me at benji.app.
Check out Imprezia at imprezia.com.
Timestamps
(00:00) – Intro
(01:00) – Building Google AdSense for AI Apps
(01:34) – Why Ads? Joining Amazon Ads in 2019
(03:11) – From Meta GenAI Ads to Starting His Own Company
(04:39) – How Fast AI Models Are Moving
(06:21) – Does Every Model Upgrade Make the Business Stronger or Weaker?
(08:02) – Do Everything Yourself Before You Hire
(09:42) – You Can’t Lead a 2,400 Elo Team at 800 Elo
(11:42) – Solving the Cold Start Problem in a Two-Sided Marketplace
(12:23) – Go-to-Market: Why Cold Email Doesn’t Work Anymore
(14:10) – Business Model: Revenue Share, No Retainer
(14:50) – Why Ad Tech Can’t Be Fully Self-Serve
(15:45) – The Future of Marketing in an AI World
(17:50) – Every Billion-User Product Runs on Ads
(18:14) – Search Is Dying: Where Ad Budgets Go Next
(19:37) – Will Google Own AI Advertising?
(21:33) – How an Ad Gets Priced: The Real-Time Auction
(23:31) – Why Context Makes Ads More Valuable
(24:40) – How CPMs Are Set
(25:41) – The Data Flywheel Behind Better Ads
(27:22) – Revenue Share and Where the Network Performs Best
(30:44) – Hiring: Hunters, Not Farmers
(32:44) – KPIs: Total Transparency Over Individual Metrics
(34:56) – The Biggest Bottleneck: Slow-Moving Advertisers
(38:08) – Will AI Kill the Big Ad Agencies?
