Koah CPO on Why Ads in AI Don't Have to Be Evil

What would it look like if every AI app had a sustainable way to make money without dark patterns or paywalls? In this podcast hosted by iDonate VP of Product and Engineering Nacho Andrade, Koah Co-Founder and CPO Mike Choi will be speaking on how he’s building the AdSense for the AI app network and why ads in AI don’t have to be evil. Mike shares what he learned from 18 months of rapid consumer app building, why AI users behave completely differently than search users, and why sponsored experiences in AI might actually be a good thing for everyone.

Subscribe to the Product Talk podcast on Spotify and Apple Podcasts and catch every conversation with leading product executives. Episodes drop on Wednesdays.

Show Notes

1. Koah is essentially AdSense for the AI app ecosystem: developers integrate an SDK to render contextually relevant sponsored results inside their AI apps, giving them a scalable way to generate revenue from users asking questions.
2. Mike and his co-founders started Koah because they felt the monetization problem firsthand. After 18 months of building one consumer app per month and struggling to generate sustainable revenue, Koah came directly from their own pain.
3. The simple “just show an ad” approach doesn’t work in AI. When Mike’s team built a basic RAG-based ad engine, they found that embedding relevance doesn’t capture the nuanced contextual relevance that actually drives real engagement and clicks.
4. AI users are in exploratory mode, not purchase mode. Unlike search, where someone typing “Nike shoes” has clear purchase intent, people asking AI assistants questions are in a top-to-middle funnel curiosity state, and ads need to be designed for that entirely different behavior.
5. The holy grail Mike is building toward is guiding users all the way from an exploratory AI query down to checkout without ever leaving the app. Getting that end-to-end journey right is the hardest and most valuable problem Koah is working on.
6. Instead of calling them ads, Koah is moving toward the term “sponsored commercial experiences.” The idea is that AI advertising can be interactive, contextual, and genuinely useful to users rather than clickbait banners interrupting a conversation.
7. Focus is the single most important discipline for a high-performing product team, and Mike admits he had to learn it the hard way. His shortcut: ask every teammate if they could only work on one thing what would it be, and then hold them to that answer.
8. Context switching is expensive, and even personal habits matter. Mike has become intentional about when he checks email and how he structures his day to reduce the cognitive cost of bouncing between tasks.
9. Building fast is one of Koah’s core superpowers, trained by 18 months of shipping one consumer app per month. That pace built a muscle for making quick decisions, launching early, and iterating based on real user feedback rather than internal assumptions.
10. Launching something that meets only 30-40% of your vision is better than waiting for perfection. Getting a product into users’ hands creates urgency to improve it, and that urgency generates its own productive flywheel.
11. Taste and design are competitive advantages at Koa. While the concept of contextual AI ads sounds simple, what the sponsored experience looks and feels like inside a conversation is where great design creates real differentiation from competitors.
12. Koah is a two-sided marketplace with two completely different user personas: developers who want technical, developer-native interfaces and advertisers who need simplicity and clarity. Designing a brand that speaks to both at the same time is genuinely difficult.
13. When prioritizing between the two sides, Mike leans toward developers and publishers first. Being strong in the technical AI surface area will eventually make Koah attractive to advertisers, so getting developer traction is the right sequencing.
14. Competition in an emerging market is actually a good thing. Every competitor that enters the AI ad space helps normalize the idea of sponsored results in AI, making it easier for everyone to convince developers and advertisers to participate.
15. Sponsored results are coming to AI, and Mike thinks they will be net positive. Just as Instagram has ads people genuinely enjoy, well-designed AI ads will increase engagement and give great brands a way to surface in front of users who are already curious and ready to explore.
16. Koah focuses on AI apps in the search and productivity space where real purchase intent exists behind queries. Apps in entertainment or relationship AI don’t fit the advertiser profile, so staying disciplined about that ICP is critical.
17. The hardest unsolved problem for Koah right now is the advertiser onboarding experience. How do you teach an advertiser how AI targeting works when you’re still figuring out what kinds of ads perform best?
18. Mike recommends the book Click, from the creators of the Design Sprint, for anyone building in the AI space. The methodology of focusing on one problem, prototyping fast, and iterating is his operating framework.
19. After a health scare, Mike started wearing a Whoop and now treats his recovery score as a real signal for how he plans his day. He goes to sleep at 8:30 when it tells him to, and says it has improved both his health and his productivity.
20. Mike’s closing message is that ads aren’t evil. They’re what allows indie developers and AI apps to build sustainable businesses, and the goal is to design sponsored experiences so contextually relevant that users actually benefit from seeing them.