---
id: "concept-bnn-vs-ann"
type: "concept"
source_timestamps: ["§ Decades of Marketing Science Built for a Different Customer"]
tags: ["marketing-science", "behavioral-economics", "persuasion", "artificial-neural-networks"]
related: ["claim-persuasion-science-gap", "concept-ai-engine-optimization"]
speakers: ["Kartik Hosanagar"]
definition: "The contrast between human consumers (BNNs) driven by psychological biases and AI agents (ANNs) driven by algorithmic weights, requiring entirely different marketing persuasion sciences."
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?"
---
# Biological vs. Artificial Neural Networks (BNNs vs. ANNs)

A conceptual frame contrasting **human consumers — Biological Neural Networks (BNNs)** — with **AI agents — Artificial Neural Networks (ANNs)** — to argue that current marketing tactics are becoming obsolete for the new buyer.

Marketers have spent decades mastering BNN persuasion through **behavioral economics, consumer psychology, and neuromarketing** — leveraging tactics like **$19.99 charm pricing, social proof, scarcity, authority, loss aversion, and specific color/layout interactions**. [[entity-kartik-hosanagar]] argues ANNs are *"like a new species"* with entirely different biases, framing effects, and decision rules (see the quote [[quote-ann-new-species]]). Because ANNs do not respond to scarcity cues or visual layouts the way BNNs do, the existing science of human persuasion does not transfer — a new science of AI-agent communication must be built (see [[claim-persuasion-science-gap]] and the action [[action-develop-ai-persuasion]]).

This is also why [[concept-ai-engine-optimization|AI Engine Optimization (AEO)]] is insufficient: being *visible* to an ANN is not the same as *persuading* one.

*Enrichment note:* conceptually sound — ANNs are driven by objective functions, training-data distributions, and algorithmic optimization, not human affect (cf. Rahwan et al.'s "machine behavior," 2019). But it is **not yet empirically demonstrated at scale**: there is no large body of published work rigorously testing charm pricing, color, or scarcity on autonomous agents. **Counter-perspective:** because agents are often trained on human behavior and optimize for human satisfaction/purchase-likelihood, human-optimized signals (strong reviews, price competitiveness) may still matter *indirectly* — so treat this as a strong hypothesis, not settled fact.


## Related across articles
- [[concept-bot-psychology-d13]]
- [[concept-bot-psychology-d29]]
- [[claim-persuasion-science-gap]]