---
id: "contrarian-poor-roi-meaning"
type: "contrarian-insight"
source_timestamps: ["¶3"]
tags: ["roi", "misconceptions"]
related: ["claim-traditional-roi-fails-ai", "concept-ai-commodity-fallacy"]
challenges: "The conventional view that low immediate ROI proves AI is an overhyped or failing investment."
sources: ["spine"]
sourceVaultSlug: "hbr-seg-spine"
originDay: 1
articleStem: "hbr-edu-47-5-types-ai-investment"
sourceUrl: "https://hbr.org/2026/06/the-5-types-of-ai-investment-and-how-to-capture-their-value"
sourceTitle: "The 5 Types of AI Investment–and How to Capture Their Value"
---
# Poor AI ROI indicates bad metrics, not bad technology

**Conventional wisdom it challenges:** that the lack of immediate financial returns on AI — as documented by [[entity-mckinsey-d1|McKinsey]], [[entity-bcg-d1|BCG]], and [[entity-deloitte-d1|Deloitte]] — proves AI investments are failing or overhyped.

**Prasad's reframe:** the poor ROI is evidence that we are using the *wrong financial instruments*, treating a highly contextual, local capability as if it were a plug-and-play commodity (the [[concept-ai-commodity-fallacy]]). The surveys measure a real thing badly; the fix is bespoke financial logic per investment type ([[claim-traditional-roi-fails-ai]], [[framework-5-types-ai-investment]]).

**Enrichment / counter-counter.** The reframe is strong, but not absolute: for narrow automation and decision-support deployments, standard ROI, payback, and total cost of ownership remain useful and sometimes necessary. "Bad metrics" is the right diagnosis for *strategic* AI; it is not a license to abandon measurement everywhere.