---
id: "action-use-transcripts-as-context"
type: "action-item"
source_timestamps: ["§ Where to Start"]
tags: ["prompt-engineering", "context-files"]
related: ["action-convene-expert-panels", "concept-codifying-judgment"]
action: "Use transcripts of expert debates as the context layer for AI agents."
outcome: "Provides agents with explicit, nuanced guidance on organizational judgment."
speakers: ["Jen Stave", "Ryan Kurt", "John Winsor"]
sources: ["agentic"]
sourceVaultSlug: "hbr-seg-agentic"
originDay: 6
articleStem: "hbr-new-27-teach-ai-your-decisions"
sourceUrl: "https://hbr.org/2026/06/teach-your-ai-how-you-make-decisions"
sourceTitle: "Teach Your AI How You Make Decisions"
---
# Use debate transcripts as AI context layers

**Action:** Use transcripts of expert debates as the context layer for AI agents.

Take the transcripts generated from expert-panel debates ([[action-convene-expert-panels]]) and use them directly as the foundational context layer for future agentic deployments. This provides the AI with explicit guidance on risk tolerance, empathy, and escalation logic — the nuance standard documentation misses.

**Outcome:** Provides agents with explicit, nuanced guidance on organizational judgment.

Depends on [[prereq-llm-context-windows]] and completes [[framework-scenario-based-extraction]]. Caveat: the open question [[question-maintaining-codified-judgment]] flags that transcripts must be maintained, and [[cp-sops-still-valuable]] warns that un-abstracted transcripts can become unwieldy.