Spec-driven development: The AI engineering workflow at Notion | Ryan Nystrom

Notion engineer Ryan Nystrom detailed the company's spec-driven development workflow, where engineers dictate ideas into OpenAI's Whisper, have Codex format them as specs, commit them to the repository, and let AI agents autonomously implement and verify the code. Nystrom, who joined Notion in December 2024 after the Campsite acquisition, also described Notion's internal "Boxy" system that lets engineers @mention Codex from within Notion comments to generate a full pull request with screenshots in 20 minutes. The approach relies on fast CI and automated context surfaces to protect engineers from burnout while enabling high-frequency, high-quality meetings.

Ryan Nystrom is a software engineer at Notion. He joined in December 2024 after Notion acquired Campsite, the team communication platform he co-founded with Brian Lovin. At Notion, he’s been a core builder of Notion AI and the Custom Agents feature launched in February 2026. He manages a team of six to seven engineers while still writing code himself, currently running Project Afterburner, a push to cut Notion’s CI time to a quarter of its current duration. What you’ll learn: How to build a Notion AI custom agent that auto-generates your daily standup pre-read by pulling from Slack, GitHub, Honeycomb metrics, and yesterday’s meeting transcript How to configure subagents and MCP integrations within Notion AI How Notion’s internal “Boxy” system lets engineers @mention Codex from within Notion comments and get a full pull request with screenshots in 20 minutes The spec-first development workflow: dictate an idea into Whisper, have Codex format it as a proper spec, commit it to the repo, and let the agent implement and verify it autonomously Why fast CI is absolutely critical in the age of AI coding agents How to prompt AI coding agents to defend their reasoning under pushback Why engineering managers and even senior executives should keep writing code Brought to you by: WorkOS —Make your app enterprise-ready today Orkes —The enterprise platform for reliable applications and agentic workflows In this episode, we cover: 00:00 https://www.youtube.com/watch?v=pUHA jNwuYE Introduction to Ryan Nystrom 02:48 https://www.youtube.com/watch?v=pUHA jNwuYE&t=168s How AI has upended 12+ years of the same working routine 04:30 https://www.youtube.com/watch?v=pUHA jNwuYE&t=270s Project Afterburner: Notion’s push to cut CI time to a quarter 09:00 https://www.youtube.com/watch?v=pUHA jNwuYE&t=540s Why high-frequency, high-quality meetings beat lower-frequency standups 11:10 https://www.youtube.com/watch?v=pUHA jNwuYE&t=670s How automated context surfaces every engineer’s work equally 12:15 https://www.youtube.com/watch?v=pUHA jNwuYE&t=735s Why cutting meeting prep is a burnout protection mechanism 14:26 https://www.youtube.com/watch?v=pUHA jNwuYE&t=866s The case for engineering managers writing code 16:13 https://www.youtube.com/watch?v=pUHA jNwuYE&t=973s Inside “Boxy”: Notion’s internal VM-based background agent system 20:30 https://www.youtube.com/watch?v=pUHA jNwuYE&t=1230s Old World vs. New World code review 24:51 https://www.youtube.com/watch?v=pUHA jNwuYE&t=1491s Prompting Codex from Notion comments 29:20 https://www.youtube.com/watch?v=pUHA jNwuYE&t=1760s The emotions around code review 31:01 https://www.youtube.com/watch?v=pUHA jNwuYE&t=1861s Quick recap 32:00 https://www.youtube.com/watch?v=pUHA jNwuYE&t=1920s Spec-first development: writing and checking agent specs into the repo 35:10 https://www.youtube.com/watch?v=pUHA jNwuYE&t=2110s The spec as changelog: version control for how a feature actually works 37:53 https://www.youtube.com/watch?v=pUHA jNwuYE&t=2273s How engineers’ roles are evolving 39:00 https://www.youtube.com/watch?v=pUHA jNwuYE&t=2340s Lightning round 45:21 https://www.youtube.com/watch?v=pUHA jNwuYE&t=2721s Where to find Ryan Tools referenced: • Notion AI: https://www.notion.com/product/ai https://www.notion.com/product/ai • Notion Custom Agents: https://www.notion.com/blog/introducing-custom-agents https://www.notion.com/blog/introducing-custom-agents • Codex OpenAI : https://openai.com/codex https://openai.com/codex • Claude Code Anthropic : https://claude.ai/code https://claude.ai/code • Honeycomb observability + MCP : https://www.honeycomb.io https://www.honeycomb.io • Whisper OpenAI voice transcription : https://openai.com/research/whisper https://openai.com/research/whisper • Slack: https://slack.com https://slack.com • GitHub: https://github.com https://github.com Other references: • How Stripe built “minions”—AI coding agents that ship 1,300 PRs weekly from Slack reactions | Steve Kaliski Stripe : https://www.chatprd.ai/how-i-ai/stripes-ai-minions-ship-1300-prs-weekly-from-a-slack-emoji https://www.chatprd.ai/how-i-ai/stripes-ai-minions-ship-1300-prs-weekly-from-a-slack-emoji • Notion 3.3 Custom Agents launch February 24, 2026 : https://www.notion.com/releases/2026-02-24 https://www.notion.com/releases/2026-02-24 Where to find Ryan Nystrom: LinkedIn: https://www.linkedin.com/in/ryannystrom/ https://www.linkedin.com/in/ryannystrom/ GitHub: https://github.com/rnystrom https://github.com/rnystrom Where to find Claire Vo: ChatPRD: https://www.chatprd.ai/ https://www.chatprd.ai/ Website: https://clairevo.com/ https://clairevo.com/ LinkedIn: https://www.linkedin.com/in/clairevo/ https://www.linkedin.com/in/clairevo/ Production and marketing by https://penname.co/ https://penname.co/ . For inquiries about sponsoring the podcast, email email protected /cdn-cgi/l/email-protection .