---
id: "concept-prompt-driven-optimization"
type: "concept"
source_timestamps: ["\\\"§ Understand the prompt", "not just the agent.\\\""]
tags: ["seo", "prompt-engineering", "consumer-research"]
related: ["action-analyze-user-prompts", "concept-ai-shopping-agents"]
definition: "The practice of optimizing product listings and data feeds based on the specific natural language instructions (prompts) consumers give to their AI shopping agents."
sources: ["geo"]
sourceVaultSlug: "hbr-seg-geo"
originDay: 3
articleStem: "hbr-tier2-06-ai-shopping-agents"
sourceUrl: "https://hbr.org/2026/05/research-traditional-marketing-doesnt-work-on-ai-shopping-agents"
sourceTitle: "Research: Traditional Marketing Doesn’t Work on AI Shopping Agents"
---
# Prompt-Driven Optimization

**Definition:** Optimizing product listings and data feeds around the specific natural-language instructions (prompts) consumers give their agents — not just around the agent's model behavior.

Unlike human shoppers, who arrive with implicit and often malleable preferences, an [[concept-ai-shopping-agents|AI shopping agent]] arrives with an **explicit mandate: the user's prompt**. The agent's behavior is entirely bounded by those instructions:

- `"find me the best-reviewed wireless headphones under £100"` triggers a fundamentally different search-and-evaluation algorithm than
- `"get me the cheapest option that ships tomorrow."`

Therefore, optimizing for agents requires **deep consumer research into how users structure their prompts** (see [[action-analyze-user-prompts]]). Brands must analyze query patterns so their product data is structured to surface favorably for the **most common and lucrative prompt structures** in their category.

> "[[quote-agent-mandate|An AI shopping agent does not arrive with its own preferences. It arrives with the user's prompt.]]"

**Enrichment context:** Agent frameworks like ACES explicitly define tasks as shopping *instructions*, and demonstrate that changing task wording ("cheapest" vs. "highest rated") alters choice behavior — validating the prompt as a first-class optimization lever. This is conceptually adjacent to SEO/GEO: orienting data structures toward likely user intents.

**Related:** [[action-analyze-user-prompts]] · [[concept-ai-shopping-agents]] · [[quote-agent-mandate]]


## Related across articles
- [[claim-query-determines-competitive-set]]
- [[quote-agent-mandate]]
- [[concept-problem-literacy]]