šļø How I AI: Quests, token leaderboards, and the elite AI adoption playbook & Notionās spec-driven development
AI Summary
This edition of How I AI features two podcast episodes: John Kim (Sendbird CEO) shares his gamified internal AI adoption playbook using quests, token leaderboards, and a skills marketplace, and Ryan Nystrom discusses spec-driven development as Notion's AI engineering workflow. Key strategies include building secure app templates for non-technical builders, measuring token usage across tiers from Beginner to AI God, and treating internal tooling as a product rather than a top-down program.
Key Facts
Author Takes
AI adoption leadership
The most important hiring criteria for AI-first companies are curiosity, agency, and energyānot tenure or experienceāand job descriptions should be rewritten accordingly.
24/7 AI utilization
The goal isn't just to use AI during work hours but to smooth token-usage curves so AI agents work around the clock, filling gaps when humans are unavailable.
Contrarian Angle
Gamified Internal AI Adoption via Quest Marketplace
John Kim built 'Automators', a gamified internal platform where any employee can post a 'quest' (automation request) and engineers or AI agents build it. Completers earn XP redeemable for gift cards or exec accessānot a mandate but a product.
Treats internal AI adoption as a consumer product with game mechanics rather than a top-down compliance program, driving organic participation.
Non-Technical Teams Shipping to Production via Pre-vetted App Templates
Sendbird created app templates with pre-configured auth, databases, and InfoSec-vetted security so marketers, salespeople, and CSMs can build and ship production apps themselves without engineering involvement.
Eliminates the engineering bottleneck for non-technical builders by pre-solving security and compliance, enabling marketing to ship a Stripe-integrated swag store with a Konami Code Easter egg in days.
More from Lenny's Newsletter
šļø This week on How I AI: Claude Cowork tutorial for non-engineers + Build your own Slack inbox (for $0)
This newsletter features two AI productivity episodes: Yash Tekriwal demonstrates building a custom AI-powered Slack inbox that reduces 150 daily noti
šļø This week on How I AI: I gave Claude Code our entire codebase. Our customers noticed.
Al Chen from Galileo shares how he uses Claude Code to query their entire codebase across 15 repositories to answer enterprise customer technical ques
šļø This week on How I AI: How Stripe built āminionsāāAI coding agents that ship 1,300 PRs per week + How to turn ā¦
This newsletter features two episodes from How I AI podcast covering AI coding agents at scale and personal productivity automation. The first episode