---
id: "claim-persuasion-science-gap"
type: "claim"
source_timestamps: ["§ Decades of Marketing Science Built for a Different Customer"]
tags: ["marketing-science", "psychology", "persuasion"]
related: ["concept-bnn-vs-ann", "action-develop-ai-persuasion"]
speakers: ["Kartik Hosanagar"]
confidence: "high"
testable: true
enrichment_status: "conceptually sound; best treated as a strong hypothesis, not settled fact"
sources: ["geo"]
sourceVaultSlug: "hbr-seg-geo"
originDay: 3
articleStem: "hbr-tier2-05-market-to-ai-customer"
sourceUrl: "https://hbr.org/2026/06/how-do-you-market-to-an-ai-customer"
sourceTitle: "How Do You Market to an AI Customer?"
---
# The science of human persuasion does not transfer to AI agents

**Claim:** Decades of marketing science — behavioral economics, consumer psychology, neuromarketing — are fundamentally incompatible with AI agents.

Tactics proven on humans ([[concept-bnn-vs-ann|BNNs]]) — pricing at **$19.99 instead of $20**, **social proof**, **artificial scarcity**, optimized **color and layout** — do not have the same effect on ANNs, which possess different biases, framing effects, and decision rules. The consequence: marketers currently **lack a scientific framework** for persuading their new primary customer, which is why [[entity-kartik-hosanagar]] calls for building one (see [[action-develop-ai-persuasion]] and the quote [[quote-ann-new-species]]).

**Confidence:** High (as stated). **Testable:** Yes.

*Enrichment status — strong hypothesis, not settled fact.*
- **Conceptually sound:** ANNs don't experience emotion, loss aversion, or visual perception; their "decisions" are algorithmic (objective functions, training data, reward signals).
- **Empirical gap:** no large published body of work rigorously tests whether classic price-ending or color tricks have *no* effect on shopping agents.
- **Counter-perspective:** agents trained on human behavior and optimizing for human satisfaction may **implicitly** learn human-centric preferences, so some tactics (strong reviews, competitive pricing) may still matter indirectly — especially in [[concept-human-present-mode|human-present]] and recommendation contexts. Frame the science as *likely-not-to-generalize* rather than *proven-incompatible*.


## Related across articles
- [[claim-traditional-marketing-fails]]
- [[concept-algorithmic-skepticism]]
- [[concept-bnn-vs-ann]]