---
id: "claim-codified-judgment-compounds"
type: "claim"
source_timestamps: ["§ A New Differentiator"]
tags: ["roi", "scaling", "compounding-returns"]
related: ["concept-judgment-infrastructure", "entity-ita-group"]
confidence: "high"
testable: true
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"
---
# Codified judgment compounds over time

**Claim (confidence: high, testable):** Building [[concept-judgment-infrastructure]] exhibits compounding returns. The first use cases — e.g., [[entity-ita-group|ITA Group]]'s first 6-7 months — are slow because the organization is learning the novel skill of translating tacit domain expertise from human heads into context files for agents.

But once this operating model takes hold, the pace of deployment and innovation accelerates dramatically. The organization builds the trust, governance, and operating rhythm required to repeat the pattern, leading to shrinking timelines (months to weeks) and the organic spread of AI-driven workflows into new functions.

**Enrichment assessment — well supported conceptually; empirically thin.** Codified policies, patterns, and governance frameworks are reusable, and AWS/HBR commentary on redesigning end-to-end value chains implies learning effects that compound. Even data-infrastructure-focused reports note that once foundations are modernized, subsequent initiatives get easier. But the evidence here is case-study and expert judgment, not longitudinal metrics — which is exactly the gap flagged by [[question-measuring-judgment-roi]].