Playbook
Tasteful Skills
“Tasteful Skills” argues that the best agent skills are not documentation or best-practice lists.
Back to Podcast Digest
Quests, token leaderboards, and a skills marketplace: the elite AI adoption playbook | John Kim
TL;DR
Sunbird treats AI adoption like an internal product, not a memo — John Kim built an internal “Automators” platform where anyone can post an AI “quest,” pair with builders, track risk and weeks saved, and now even have AI generate PRDs and start coding from the spec.
The most vivid proof point was a marketer-built swag store shipped in 1-2 days — Sunbird’s marketing team launched a Stripe-integrated store called “Big S Energy” with jokes, Easter eggs like the Konami code, and event promotion, all without engineering support.
They gamified usage with a token leaderboard that ranks employees from beginner to ‘AI god’ — Managers see team-level token usage, Sunbird defines “AI god” as 100 million+ tokens per day, and John says the goal isn’t performance review but knowing how to coach each person to the next level.
Sunbird created a company-wide skills marketplace so expertise becomes reusable software — Teams can publish plugins and markdown-based “skills” for sales, recruiting, design, and more, reducing duplicated effort and letting people plug internal know-how into their own workflows.
The adoption playbook depends on infrastructure plus incentives, not just enthusiasm — Sunbird built secure app templates, internal Git and deployment guides, and an AI internal-ops team reporting to John and the chief of staff, while rewarding quest builders with XP, gift cards, executive time, and company-wide demos.
John’s hiring thesis has shifted toward curiosity, agency, and energy over tenure — He says AI lowers the cost of learning and building to roughly a $20-$200/month tool budget, so the edge now goes to people who will “fail forward,” teach themselves fast, and make the most of always-on AI partners.
The Breakdown
The opener: AI for marketers, not just engineers
John starts with the big idea: becoming “AI first” means making AI part of the workforce, not just another productivity tool. Clarvo immediately frames what makes Sunbird interesting — they’ve turned adoption into a product experience for employees, with systems, incentives, and visibility.
A swag store that proves marketers can now ship real software
The first demo is delightfully unserious in the best way: Sunbird’s marketing team built a full swag store called “Big S Energy,” complete with Stripe payments, joke merch like “my AAS is bigger than your SaaS,” and a Konami code Easter egg that reveals info about their May 7 Delight Spark conference in San Francisco. The point lands hard: in the old world, this would have died in roadmap triage; with AI, marketers can ship customer-facing experiences in a day or two.
The “quest” system: an internal marketplace for AI work
To get non-engineers building, Sunbird created an internal platform called Automators, where anyone can post a “quest” for an automation or tool. A finance person can request help with accounts receivable, a recruiter can ask for a workflow, and engineers or AI-enabled teammates can jump in, collaborate, and return a code repo or reusable skill.
What makes quests work: templates, safety rails, and dopamine
Clarvo calls out the clever mechanics: each quest can track things like risk, time saved, and who benefits, which turns vague AI enthusiasm into concrete internal ROI. Behind the scenes, builders earn XP redeemable for gift cards, executive coffee chats, and company-wide Wednesday demos — and John says that instant feedback loop gives people a real dopamine hit compared with waiting through sprint planning.
AI now joins the builders, not just the requesters
Sunbird is rolling out the next layer: AI can read a quest spec, draft a PRD, and start coding. That’s paired with internal guides for GitHub, deployment, and secure app-building, plus prebuilt templates with authentication, compliance, and infrastructure already handled so a marketer or CSM can focus on the idea instead of plumbing.
The skills marketplace: turning tribal knowledge into reusable assets
From there, John shows a company-wide plugin and skills marketplace where teams upload reusable markdown-based expertise. Sales can share MEDDIC advisors, recruiting and design can publish their own patterns, and the whole goal is to stop different teams from quietly rebuilding the same app or workflow in silos.
Marketing’s internal SaaS stack and the rise of custom micro-software
John then shows how far this has gone inside marketing: they’ve built an internal suite for interview planning, account-based marketing, competitor review, campaign tracking, and a “buzzboard” for social campaigns. One live example supports Sunbird’s San Francisco billboard campaign, letting employees pick images, tweak copy and tone, and post directly to LinkedIn while tracking who drives the most engagement.
Measuring adoption without shame: the token leaderboard
One of the strongest moments is John’s dashboard for token usage across the company, teams, and individuals. Sunbird classifies people into five tiers from beginner to catalyst to “AI god” — defined as 100 million+ tokens a day — and uses that not for formal performance review, but to understand who needs help and what AI-first looks like by function.
Personal AI, gamer energy, and the hiring shift
In the closing stretch, John shares personal projects like “the gardener,” an open-source tool that tends his Obsidian notes by enriching entities, fixing structure, and cross-linking ideas, plus custom offline learning hubs for neuroscience, quantum mechanics, and fusion. He says Sunbird has rewritten job descriptions to lower the emphasis on tenure and raise the premium on curiosity, agency, and energy — and he ends with a founder’s playbook: find the naturally curious people already experimenting, spotlight them, and let that energy spread.
Share
Keep Reading
Two Ways to Follow
The Weekly Echo. The inbox-shaped summary of what mattered.
New editorials announced here.
