---
id: "action-run-ai-experiments"
type: "action-item"
source_timestamps: ["§ Controlled Experimentation"]
tags: ["measurement", "a-b-testing"]
related: ["concept-controlled-experimentation-ai"]
action: "Design and execute controlled A/B experiments to measure Gen AI productivity impacts."
outcome: "Empirical proof of AI value and capability building within internal data science teams."
speakers: ["Tom Davenport", "John J. Sviokla"]
sources: ["spine"]
sourceVaultSlug: "hbr-seg-spine"
originDay: 1
articleStem: "hbr-cl-95-6-disciplines-genai"
sourceUrl: "https://hbr.org/2024/07/the-6-disciplines-companies-need-to-get-the-most-out-of-gen-ai"
sourceTitle: "The 6 Disciplines Companies Need to Get the Most Out of Gen AI"
---
# Run Controlled AI Experiments

**Action:** Design and execute controlled A/B experiments to measure Gen AI productivity impacts.

**How:** Design controlled experiments comparing groups using Gen AI against those who are not, to statistically determine actual productivity and effectiveness gains in specific domains. Also test modalities (solo generator vs. co-pilot). Requires the background in [[prereq-ab-testing-stats]].

**Expected outcome:** Empirical proof of AI value *and* capability building within internal data-science teams (rather than outsourcing measurement to academics/vendors).

Operationalizes the discipline [[concept-controlled-experimentation-ai]] and feeds [[concept-business-value-measurement]].