Skip to content
How I AIHow I AI

Stripe built a company brain: Meet Kai

Sharadh Krishnamurthy is an engineering manager at Stripe, where he helped build Kai, the company’s internal AI agent used by more than 10,000 employees every week. He’s worked across several of Stripe’s core infrastructure teams, including data and developer experience, which gives him a grounded, systems-level perspective on what it actually takes to make AI work at enterprise scale. He’s currently focused on the governance, skills, and infrastructure layers that let every Stripe employee use AI safely and effectively, regardless of their technical background. *What you’ll learn:* 1. Why Stripe built Kai from scratch instead of buying, and what tipped the decision 2. What Kai knows about you by default and what you actually control 3. Why “projects” at Stripe are a governance mechanism, not just a folder 4. How Stripe structured its data layer so agents can query safely at scale 5. Why the infrastructure Stripe built for human developers turned out to be exactly what agents needed 6. How Kai’s skills platform lets any employee package a workflow, and what happens when you have 2,000 of them 7. What Sharadh learned the hard way when agents nearly took down production systems *Brought to you by:* DX—Engineering intelligence for the AI era: https://getdx.com/howiai Hyperagent—Deploy fleets of agents that handle real work: https://www.hyperagent.com/howiai *In this episode, we cover:* (00:00) Introducing Sharadh (02:46) Why Stripe built an AI agent (Kai) instead of buying tools (05:18) What Kai knows about you (and what you can turn off) (06:51) Projects as a governance layer (10:04) Live demo: Kai builds a dashboard (12:18) Tools, skills, and the secure sandbox (17:22) Why Stripe has benefited so much from AI (19:20) Agentic identity, load shedding, and rogue agents (20:41) Iterating on the dashboard (25:01) How they rolled out Kai across the team (29:07) How projects work (34:18) Bespoke agents for bespoke use cases (35:58) The skill builder workflow (40:40) Skill quality, evals, and telemetry (43:01) Recap (45:13) Lightning round *Blog and detailed workflow walkthroughs from this episode:* How Stripe Built Kai: Data Dashboards, Reusable Skills, and Enterprise AI Governance: https://www.chatprd.ai/how-i-ai/how-stripe-built-kai-data-dashboards-reusable-skills-and-enterprise-ai-governance ↳ How to Create a Custom Data Dashboard on the Fly with a Natural Language AI Agent: https://www.chatprd.ai/how-i-ai/workflows/how-to-create-a-custom-data-dashboard-on-the-fly-with-a-natural-language-ai-agent ↳ How to Turn a One-Off AI Chat Session into a Reusable Workflow for Your Team: https://www.chatprd.ai/how-i-ai/workflows/how-to-turn-a-one-off-ai-chat-session-into-a-reusable-workflow-for-your-team ↳ How to Implement AI Governance with Context-Aware Project Controls: https://www.chatprd.ai/how-i-ai/workflows/how-to-implement-ai-governance-with-context-aware-project-controls *Tools referenced:* • Trino: https://trino.io/ • Anthropic: https://www.anthropic.com/ • Gemini: https://gemini.google.com/ • Cursor: https://www.cursor.com/ *Where to find Sharadh Krishnamurthy:* LinkedIn: https://www.linkedin.com/in/sharadhk *Where to find Claire Vo:* ChatPRD: https://www.chatprd.ai/ Website: https://clairevo.com/ LinkedIn: https://www.linkedin.com/in/clairevo/ X: https://x.com/clairevo _Production and marketing by https://penname.co/._ _For inquiries about sponsoring the podcast, email jordan@penname.co._

Sharadh KrishnamurthyguestClaire Vohost
Sep 7, 202650mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

Stripe’s Kai: a governed, context-aware company brain with projects and skills

  1. Stripe built Kai to scale AI to the entire company with the right governance structures, rather than relying on off-the-shelf tools that don’t match Stripe’s complexity and security needs.
  2. Kai is context-aware (identity and org chart by default) and can optionally connect to sensitive systems like Drive and Slack with employee-controlled access toggles.
  3. “Projects” are the central mechanism for enterprise governance, defining intent, tool policies, human-in-the-loop approvals, and model/cost constraints for specific teams or workflows.
  4. A live demo shows Kai creating and then iterating on a data dashboard using internal skills/tools plus a secure sandbox, highlighting how agents can help non-engineers build “light apps.”
  5. Stripe attributes strong AI leverage to pre-existing platform investments (DevX and data/analytics foundations) and emphasizes infra hardening to prevent agent-driven failure modes (brute force queries, rogue actions).

IDEAS WORTH REMEMBERING

5 ideas

Governance—not model quality—was the core reason Stripe built Kai.

Stripe didn’t just want “AI access”; it needed enterprise-grade guardrails so employees can use agents safely without thinking about permissions, models, or connectors on every task.

Personalization is opt-in and user-controlled, not an invisible data grab.

Kai starts with basic identity/org-chart context, then can optionally connect to sensitive sources like Drive and Slack. Employees can granularly control and toggle what Kai can see, even session-by-session.

“Projects” are a governance and configuration layer, not just a folder for chats.

Projects act as a shared container for intent + configuration: default models (cost/latency), which skills/tools are in scope, and which actions require confirmation. This enables different safety and capability profiles for different teams (e.g., HR).

Tools + skills + a secure sandbox turn chat into repeatable work execution.

Kai uses tools (capabilities) and skills (packaged workflows over tools) plus a secure cloud sandbox so it can write/run code and manipulate data without running on the user’s laptop or leaking across sessions.

Reliable data agents require a tiered retrieval strategy and resilient data infra.

Stripe’s “Ask Data” approach routes agents through progressively lower-level sources: existing reports/artifacts → blessed analytics layer → data catalog/SQL. This reduces incorrect queries and protects the warehouse from brute-force agent behavior.

WORDS WORTH SAVING

5 quotes

Agents are very creative at bringing your infra down.

Sharadh Krishnamurthy

It turns out that agents just, like, dial up all your failure modes. Like, it just- It just multiplies the amplitude of problems you can get, right?

Sharadh Krishnamurthy

If you put too much friction in front of people, they're just gonna do unsafe things because that's how humans are, right?

Sharadh Krishnamurthy

Your data warehouse has to be very resilient to high volume queries because when in doubt, an agent will just brute force it.

Claire Vo

Double the size of your DevX team. Double the size of your data team.

Claire Vo

Rationale for building vs. buyingContext awareness and user-controlled connectorsProjects as governance: model routing, tool policies, approvalsSkills library: creation, sharing, routingSecure cloud sandbox for code/data workData agent readiness: analytics layer, catalog, warehouse resilienceAgent risk: rogue behavior, load shedding, agentic identity

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.