GTM AI Podcast with Coach K and Jonathan Moss This AI Agent Builds Account Plans in 90 Seconds (Here's How)
Mar 24, 2026
Justin Driesse, Director of Sales Enablement at Legora and former enablement lead at Amazon, Twitch, and Slack, built a Notion AI agent that generates account plans in 90 seconds. He explains the five-prompt workflow, why Notion makes a powerful RAG system, and how agentic processes compress learning, automate SKO personalization, and surface deal patterns across accounts.
AI Snips
Chapters
Transcript
Episode notes
Account Plans Generated In 90 Seconds
- Legora replaced months of account-planning change management with a Notion agent that builds full plans in ~90 seconds when a rep types the account name and says "yes chef".
- The agent outputs tiered stakeholder maps, competitive intel, and inline footnoted sources, and Justin's CRO reacted, "Is this real?".
Notion As The Ultimate RAG And Agent Layer
- Notion functions as an inverted agent platform where one agent plus many persona/instruction pages replace many separate agents, making complex prompts a human-friendly document.
- This makes Notion the RAG (retrieval) layer and lets teams update knowledge pages to change agent behavior instantly.
Specify Allowed Sources And Exclusions
- Be explicit about what the agent should use and twice as explicit about what it must avoid to prevent stale or diluted data from Slack or other sources.
- Justin enforces guardrails by listing allowed product pages and excluding ambiguous sources in prompts.