Notion’s Token Town: 5 Rebuilds, 100+ Tools, MCP vs CLIs and the Software Factory Future — Simon Last & Sarah Sachs of Notion
April 15, 2026
AI Summary
5 min readNotion has been building AI agents since late 2022, when the team first got access to GPT-4 and tried to make an "assistant" that could use all of Notion's tools and run work in the background. It failed repeatedly. "The models were just too dumb and the context window was also way too short," Simon Last recalls. The real unlock came with Sonnet 3.5 in early 2024, and the product that shipped — custom agents that can be given permissions, set up via chat, and run in the background — became Notion's most successful launch ever in terms of free trials and conversions. But the path from 2022 to that launch involved rebuilding the agent harness five times, learning hard lessons about what models actually want, and building a team culture comfortable deleting its own code.
The five rebuilds: learning what the model wants
Continue reading the full summary in the app — free to try.
Read Full Summary →Free • No credit card required
Never miss an episode of Latent Space: The AI Engineer Podcast
Get every new episode summarized in your inbox — free, ~5 minutes to read.
No spam. Unsubscribe anytime.
What you'll learn
- 1 Notion’s Token Town: 5 Rebuilds, 100+ Tools, MCP vs CLIs and the Software Factory Future
- 2 (00:00) **MCP vs CLI Debate** - Simon declares bullishness on CLIs while still seeing MCP's role for narrow, permissioned agents
- 3 (00:24) **Notion's Commitment to MCP** - Sarah explains why Notion will support MCP regardless of internal preferences
- 4 (00:50) **Custom Agents Launch Retrospective** - The team reflects on their most successful launch ever in terms of free trials and conversions
- 5 (01:53) **Three Years of Trying Before It Worked** - Simon traces the journey from GPT-4 access in late 2022 to shipping custom agents
- 6 (04:18) **The Portfolio Approach to AI** - How Notion balances shipping working products with "AGI pilled" projects
- 7 (05:53) **Two Critical Skills for Frontier Capabilities** - Sarah explains what sets Notion apart in knowing when to push and when to wait
+ Full timestamped outline available in the app
Guests on this episode
Show Notes
For all those who missed out on London, see you in Miami next week!
Notion, the knowledge work decacorn, has been building AI tooling since before ChatGPT, with many hits from Q&A in 2023 and unified AI in 2024 and Meeting Notes in 2025. At the end of their last Make user conference, Ryan Nystrom teased Notion 3.0’s Custom Agents - and they are finally embracing the Agent Lab playbook!
Sarah Sachs and Simon Last of Notion join us for a deep dive into how Notion built Custom Agents, why it took years and multiple rebuilds to get right, and what it means to turn a productivity tool into an agent-native system of record for enterprise work.
We go inside the product, engineering, evals, pricing, and org design decisions behind one of the most ambitious AI product efforts in software today — from early failed tool-calling experiments in 2022 to agent harnesses, progressive tool disclosure, meeting notes as data capture, and the long-term vision for software factories and agentic work.
We discuss:
* Sarah and Simon’s path to launching Notion Custom Agents, and why the feature was rebuilt four or five times before it was ready for production
* Why early agent attempts failed: no tool-calling standard, short context windows, unreliable models, and too much complexity exposed to the model
* The “Agent Lab” thesis: not just wrapping a model, but understanding how people collaborate and building the right product system around frontier capabilities
* How Notion thinks about roadmap timing: not swimming upstream against model limitations, but also building early enough that the product is ready when the models are
* Why coding agents feel like the kernel of AGI, and how Notion is thinking about “software factories” made up of agents that spec, code, test, debug, review, and maintain codebases together
* How Sarah runs AI engineering at Notion (“ More from this podcast