---
type: "synthesis"
arc: "bot-behavior"
articles: ["a005", "a006", "a013", "a025", "a029"]
tags: ["persuasion-penalty", "bot-behavior", "marketing"]
id: "cross-day-persuasion-penalty-convergence"
sources: ["geo"]
sourceVaultSlug: "hbr-seg-geo"
originDay: 3
articleStem: "hbr-seg-geo"
sourceUrl: "(unified vault: 13 sources)"
sourceTitle: "HBR — Demand Ⅰ-A · GEO / AI-mediated discovery & agentic commerce"
---
The corpus's most striking empirical convergence: **five separate research teams, using different methods, all find that advanced AI actively down-weights or penalizes overt persuasion.** This is stronger than 'AI ignores marketing' — the endpoint is *skepticism*, not indifference.

- **Sabbah & Acar** (16,000 choices, 4 models): reasoning models penalize scarcity/strike-through cues — [[concept-algorithmic-skepticism]], [[claim-traditional-marketing-fails]], [[quote-persuasion-penalty]]. Only price and ratings survive ([[claim-ratings-and-price-are-universal]]).
- **Puntoni/Hermann/Schweidel** (Columbia/Yale sandbox): agents penalize 'sponsored' tags and reward organic endorsements — [[claim-sponsored-penalty]]; they are *more rational than humans* on ads ([[contrarian-bot-rationality]]).
- **Gale/Cian/Wathieu**: 78.7% of AI mentions are already positive, so sentiment-gaming is wasted — [[claim-inclusion-is-bottleneck]], [[contrarian-sentiment-optimization]].
- **Dubois/Hess/Dawson/Jaiswal (luxury)**: implicit prestige cues and white space are *penalized* — [[claim-ai-ignores-implicit-cues]], [[contrarian-white-space-penalty]].
- **Hosanagar**: decades of BNN persuasion science don't transfer to ANNs — [[claim-persuasion-science-gap]], [[concept-bnn-vs-ann]].

The unifying prescription is Sabbah & Acar's blunt one: sometimes *dial persuasion back* ([[quote-dial-it-back]]). The shared caveat across all five: the *systematic* penalty is a strongly-supported hypothesis, not a proven law, and can flip by category/model (see [[cross-day-model-heterogeneity]]). This is the behavioral engine behind the whole [[cross-day-bot-psychology-discipline]].