--- title: Notion's Token Town — Sarah Sachs, Notion | daily.dev description: Sarah Sachs, head of AI engineering at Notion, shares strategies for managing AI token costs sustainably. Key insights include staying model-agnostic to... canonical: https://daily.dev/posts/notion-s-token-town-sarah-sachs-notion-lholbr96w twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Notion's Token Town — Sarah Sachs, Notion | daily.dev og:description: Sarah Sachs, head of AI engineering at Notion, shares strategies for managing AI token costs sustainably. Key insights include staying model-agnostic to... og:url: https://daily.dev/posts/notion-s-token-town-sarah-sachs-notion-lholbr96w og:image: https://api.daily.dev/og/posts/LHolBR96w.png og:image:alt: Notion's Token Town — Sarah Sachs, Notion og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Notion's Token Town — Sarah Sachs, Notion **[AI Engineer](https://daily.dev/sources/aidotengineer)** · 23 min read · 1 upvotes · 0 comments ## Summary Sarah Sachs, head of AI engineering at Notion, shares strategies for managing AI token costs sustainably. Key insights include staying model-agnostic to maintain negotiation leverage, recognizing that AI model suppliers are often direct competitors, routing different traffic types to appropriately-priced models rather than always using the most capable (and expensive) ones, leveraging open-weight models for moderate tasks, and using CPUs instead of GPUs for deterministic workloads. She also covers multi-agent orchestration challenges, the 'lethal trifecta' security risk in agentic systems, and demonstrates Notion's own software factory workflow using Claude, Codex, and partner agents collaboratively. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is the lethal trifecta in AI agent security? The lethal trifecta refers to a dangerous combination in autonomous AI systems: access to private data, exposure to untrusted content (through ingestion, MCP, or email), and the ability to communicate externally, including payloads embedded in web search results. When a system has all three simultaneously, it creates significant unsupervised risk, and the more autonomous the system, the greater the exposure. _Teams building autonomous agents can use daily.dev to keep up with emerging security frameworks like this one._ ### Why is relying on a single AI model provider risky for a product's cost structure? Building an AI product tied to one frontier lab creates a bad deal because that lab is effectively also a competitor selling the same tokens first-party at a much lower cost of goods, then reselling to you at a markup which you resell again. Without the ability to switch providers when pricing changes unfairly, a company loses negotiating leverage and risks its business viability entirely. _Developers weighing AI vendor commitments can track pricing and model shifts on daily.dev before locking in._ ### What percentage of Notion's AI traffic is handled by an auto-selecting model instead of the most expensive frontier model? Notion's auto model handles about 75% of its AI traffic, routing tasks to whichever model fits the capability and cost needs rather than defaulting to the priciest frontier option. This model-agnostic routing lets Notion offer state-of-the-art models when needed while avoiding vendor lock-in and unnecessary token spend on simple tasks like triaging an email inbox. _Engineers designing multi-model routing systems can find similar cost-optimization approaches on daily.dev._ ## Similar posts on daily.dev - [Local AI Weekly \#3: The Token Rug Pull](https://daily.dev/posts/local-ai-weekly-3-the-token-rug-pull-h4ipcuev9) · It's Foss · 1 upvotes · 0 comments - [What Is Tokenomics, And Why Your AI Infrastructure Is Now a FinOps Problem](https://daily.dev/posts/what-is-tokenomics-and-why-your-ai-infrastructure-is-now-a-finops-problem-yrbmaidqg) · Cast AI · 0 upvotes · 0 comments - [15 Techniques to lower the cost of AI](https://daily.dev/posts/15-techniques-to-lower-the-cost-of-ai-zhqxndu8c) · Lightbend · 1 upvotes · 0 comments - [Agentic AI: How to Save on Tokens](https://daily.dev/posts/agentic-ai-how-to-save-on-tokens-vjqsrml8b) · Towards Data Science · 2 upvotes · 0 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#ai-agents](https://daily.dev/tags/ai-agents), [#notion](https://daily.dev/tags/notion) [View this post on daily.dev](https://daily.dev/posts/notion-s-token-town-sarah-sachs-notion-lholbr96w) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Notion's Token Town — Sarah Sachs, Notion","url":"https://daily.dev/posts/notion-s-token-town-sarah-sachs-notion-lholbr96w","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/notion-s-token-town-sarah-sachs-notion-lholbr96w"},"datePublished":"2026-07-23T17:55:26.549Z","dateModified":"2026-09-13T19:42:03.010Z","description":"Sarah Sachs, head of AI engineering at Notion, shares strategies for managing AI token costs sustainably. Key insights include staying model-agnostic to...","image":"https://i.ytimg.com/vi/-I5W5QVAT8E/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/-I5W5QVAT8E/sddefault.jpg","isAccessibleForFree":true,"articleSection":"AI Engineer","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"AI Engineer","logo":"https://media.daily.dev/image/upload/s--u5PucxNT--/f_auto/v1724338940/logos/aidotengineer","url":"https://daily.dev/sources/aidotengineer"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/notion-s-token-town-sarah-sachs-notion-lholbr96w","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,ai-agents,notion","timeRequired":"PT23M","video":{"@type":"VideoObject","name":"Notion's Token Town — Sarah Sachs, Notion","description":"Sarah Sachs, head of AI engineering at Notion, shares strategies for managing AI token costs sustainably. Key insights include staying model-agnostic to...","thumbnailUrl":"https://i.ytimg.com/vi/-I5W5QVAT8E/sddefault.jpg","uploadDate":"2026-07-23T17:55:26.549Z","duration":"PT23M","url":"https://api.daily.dev/r/LHolBR96w","embedUrl":"https://www.youtube.com/embed/-I5W5QVAT8E"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"AI Engineer","item":"https://daily.dev/sources/aidotengineer"},{"@type":"ListItem","position":3,"name":"Notion's Token Town — Sarah Sachs, Notion"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/notion-s-token-town-sarah-sachs-notion-lholbr96w#faq","mainEntity":[{"@type":"Question","name":"What is the lethal trifecta in AI agent security?","acceptedAnswer":{"@type":"Answer","text":"The lethal trifecta refers to a dangerous combination in autonomous AI systems: access to private data, exposure to untrusted content (through ingestion, MCP, or email), and the ability to communicate externally, including payloads embedded in web search results. When a system has all three simultaneously, it creates significant unsupervised risk, and the more autonomous the system, the greater the exposure. Teams building autonomous agents can use daily.dev to keep up with emerging security frameworks like this one."}},{"@type":"Question","name":"Why is relying on a single AI model provider risky for a product's cost structure?","acceptedAnswer":{"@type":"Answer","text":"Building an AI product tied to one frontier lab creates a bad deal because that lab is effectively also a competitor selling the same tokens first-party at a much lower cost of goods, then reselling to you at a markup which you resell again. Without the ability to switch providers when pricing changes unfairly, a company loses negotiating leverage and risks its business viability entirely. Developers weighing AI vendor commitments can track pricing and model shifts on daily.dev before locking in."}},{"@type":"Question","name":"What percentage of Notion's AI traffic is handled by an auto-selecting model instead of the most expensive frontier model?","acceptedAnswer":{"@type":"Answer","text":"Notion's auto model handles about 75% of its AI traffic, routing tasks to whichever model fits the capability and cost needs rather than defaulting to the priciest frontier option. This model-agnostic routing lets Notion offer state-of-the-art models when needed while avoiding vendor lock-in and unnecessary token spend on simple tasks like triaging an email inbox. Engineers designing multi-model routing systems can find similar cost-optimization approaches on daily.dev."}}]} ```