--- schema: 1 type: video title: AI in GTM at Notion — Flora Liu description: >- The champion just left. Do not contact this customer again. Blocked in legal. The facts that decide what a rep does next usually sit in someone's meeting notes, and Flora Liu's point is that an automation which cannot read them will eventually do something catastrophically wrong. Liu, an engineer on Notion's GTM engineering team, says a year ago this looked like a marketing ops problem and now looks like one of the more interesting distributed systems problems they have worked on. The starting state was a spiderweb: customer data across Salesforce, Gong and Outreach, product usage in Snowflake, a decade of context in docs, and every department quietly building its own agents in single player mode. The team reduced every workflow to four questions: what do we know, what should happen next, how do we execute that safely, and did it work. Those became four layers, held together by one rule: humans and agents operate on the same substrate, not an AI layer bolted on top. Snowflake computes the truth, DynamoDB serves it as a denormalized profile agents query in milliseconds, and it lands back in Notion where reps already work. Signals turn customer events into tasks, including external ones like a funding round or a shift in tech stack, and each becomes a durable Temporal workflow so one malformed transcript cannot take down a batch. Agents never speak to customers; a contact sales form counts as untrusted input. Thirteen weeks in, enterprise reps log more qualified opportunities, and users given context aware recommendations were 63% more likely to take the next step. Speaker info: - https://twitter.com/floppyliu - https://www.linkedin.com/in/flofloliu/ - https://www.flofloliu.com/ Timestamps: 0:00 - From marketing ops problem to distributed systems problem 2:23 - One journey for the customer, disconnected systems inside 4:34 - Data quality, latency, and the notes nobody can parse 5:36 - Four questions that became four layers 6:39 - Humans and agents on the same loop 7:29 - Agents never talk to the customer 9:01 - Snowflake computes truth, DynamoDB serves it 11:20 - Signals: turning customer events into tasks 13:04 - Shadowing the best reps, encoded on Temporal 16:04 - A rep's day starting from a prioritized task box 17:21 - Build or buy, decided per layer 18:52 - Early numbers and takeaways canonical: >- https://liquidgold.cc/channel/aidotengineer/explore/video/ai-in-gtm-at-notion-flora-liu-52f06d1b5314 markdown: >- https://liquidgold.cc/channel/aidotengineer/video/ai-in-gtm-at-notion-flora-liu-52f06d1b5314.md llms: null parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-24T04:45:01.221Z' language: null children: [] videoId: L4I7WgiEquo durationSeconds: 1275 publishedAt: '2026-08-26T14:00:05.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 25 endSeconds: 149 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=25' heldReason: rights - ordinal: 1 text: null startSeconds: 25 endSeconds: 149 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=25' heldReason: rights - ordinal: 2 text: null startSeconds: 149 endSeconds: 159 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=149' heldReason: rights - ordinal: 3 text: null startSeconds: 159 endSeconds: 274 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=159' heldReason: rights - ordinal: 4 text: null startSeconds: 159 endSeconds: 274 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=159' heldReason: rights - ordinal: 5 text: null startSeconds: 274 endSeconds: 286 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=274' heldReason: rights - ordinal: 6 text: null startSeconds: 286 endSeconds: 336 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=286' heldReason: rights - ordinal: 7 text: null startSeconds: 286 endSeconds: 336 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=286' heldReason: rights - ordinal: 8 text: null startSeconds: 348 endSeconds: 406 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=348' heldReason: rights - ordinal: 9 text: null startSeconds: 406 endSeconds: 416 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=406' heldReason: rights - ordinal: 10 text: null startSeconds: 449 endSeconds: 459 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=449' heldReason: rights - ordinal: 11 text: null startSeconds: 459 endSeconds: 549 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=459' heldReason: rights - ordinal: 12 text: null startSeconds: 459 endSeconds: 549 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=459' heldReason: rights - ordinal: 13 text: null startSeconds: 549 endSeconds: 561 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=549' heldReason: rights - ordinal: 14 text: null startSeconds: 561 endSeconds: 685 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=561' heldReason: rights - ordinal: 15 text: null startSeconds: 685 endSeconds: 696 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=685' heldReason: rights - ordinal: 16 text: null startSeconds: 696 endSeconds: 787 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=696' heldReason: rights - ordinal: 17 text: null startSeconds: 787 endSeconds: 797 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=787' heldReason: rights - ordinal: 18 text: null startSeconds: 797 endSeconds: 969 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=797' heldReason: rights - ordinal: 19 text: null startSeconds: 797 endSeconds: 969 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=797' heldReason: rights - ordinal: 20 text: null startSeconds: 980 endSeconds: 1047 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=980' heldReason: rights - ordinal: 21 text: null startSeconds: 980 endSeconds: 1047 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=980' heldReason: rights - ordinal: 22 text: null startSeconds: 1058 endSeconds: 1138 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=1058' heldReason: rights - ordinal: 23 text: null startSeconds: 1138 endSeconds: 1149 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=1138' heldReason: rights - ordinal: 24 text: null startSeconds: 1149 endSeconds: 1275 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=1149' heldReason: rights - ordinal: 25 text: null startSeconds: 1149 endSeconds: 1275 watch: 'https://liquidgold.cc/watch/L4I7WgiEquo?t=1149' heldReason: rights --- # AI in GTM at Notion — Flora Liu The champion just left. Do not contact this customer again. Blocked in legal. The facts that decide what a rep does next usually sit in someone's meeting notes, and Flora Liu's point is that an automation which cannot read them will eventually do something catastrophically wrong. Liu, an engineer on Notion's GTM engineering team, says a year ago this looked like a marketing ops problem and now looks like one of the more interesting distributed systems problems they have worked on. The starting state was a spiderweb: customer data across Salesforce, Gong and Outreach, product usage in Snowflake, a decade of context in docs, and every department quietly building its own agents in single player mode. The team reduced every workflow to four questions: what do we know, what should happen next, how do we execute that safely, and did it work. Those became four layers, held together by one rule: humans and agents operate on the same substrate, not an AI layer bolted on top. Snowflake computes the truth, DynamoDB serves it as a denormalized profile agents query in milliseconds, and it lands back in Notion where reps already work. Signals turn customer events into tasks, including external ones like a funding round or a shift in tech stack, and each becomes a durable Temporal workflow so one malformed transcript cannot take down a batch. Agents never speak to customers; a contact sales form counts as untrusted input. Thirteen weeks in, enterprise reps log more qualified opportunities, and users given context aware recommendations were 63% more likely to take the next step. Speaker info: - https://twitter.com/floppyliu - https://www.linkedin.com/in/flofloliu/ - https://www.flofloliu.com/ Timestamps: 0:00 - From marketing ops problem to distributed systems problem 2:23 - One journey for the customer, disconnected systems inside 4:34 - Data quality, latency, and the notes nobody can parse 5:36 - Four questions that became four layers 6:39 - Humans and agents on the same loop 7:29 - Agents never talk to the customer 9:01 - Snowflake computes truth, DynamoDB serves it 11:20 - Signals: turning customer events into tasks 13:04 - Shadowing the best reps, encoded on Temporal 16:04 - A rep's day starting from a prioritized task box 17:21 - Build or buy, decided per layer 18:52 - Early numbers and takeaways Canonical: https://liquidgold.cc/channel/aidotengineer/explore/video/ai-in-gtm-at-notion-flora-liu-52f06d1b5314 · Channel: [AI Engineer](https://liquidgold.cc/channel/aidotengineer.md) > Search chunk text is held until public-source provenance is verified. ## Document chunks > Missing: no stored document chunks are available.