---
id: "concept-share-of-model"
type: "concept"
source_timestamps: ["¶2"]
tags: ["marketing-metrics", "llm-optimization", "brand-perception"]
related: ["action-monitor-share-of-model", "entity-pernod-ricard", "concept-prompt-based-optimization", "entity-jellyfish", "prereq-seo-mechanics"]
definition: "The measure of how frequently and favorably a brand appears in generative AI and LLM outputs relative to competitors."
sources: ["agentic"]
sourceVaultSlug: "hbr-seg-agentic"
originDay: 6
articleStem: "hbr-ext-18-preparing-brand-agentic-ai"
sourceUrl: "https://hbr.org/2026/03/preparing-your-brand-for-agentic-ai"
sourceTitle: "Preparing Your Brand for Agentic AI"
---
# Share of Model

**Share of Model** is a novel marketing metric: how often and how favorably a brand shows up in AI model results compared with its competitors. In the agentic era it replaces the older lenses of 'share of voice' and 'share of search' (contrast with [[prereq-seo-mechanics-d6]]).

[[entity-pernod-ricard-d6]] pioneered actively managing this metric after discovering that LLMs were miscategorizing its affordable, mass-market Ballantine's Scotch whiskey as a *prestige* product — and partnered with [[entity-jellyfish-d6]] to iteratively prompt and correct model perceptions. Managing share of model is operational, not one-off: teams regularly prompt popular LLMs with questions about their catalog, audit the responses, and iteratively update website and advertising copy so the models ingest and echo the correct brand messaging (see [[action-monitor-share-of-model]] and the ongoing practice of [[concept-prompt-based-optimization]]).

**Enrichment / verification.** Multiple practitioner sources corroborate share of model as the AI-era analogue to share of voice, but it remains an *emerging, practitioner-defined* concept rather than a ratified industry standard. Its closest formal adjacents are **Generative Engine Optimization (GEO)** and **Answer Engine Optimization (AEO)**. The stronger causal claim that copy edits can 'force' models to echo messaging is overstated — evidence supports influence/optimization, not deterministic control. A counter-view holds the metric is still too unstable across models, prompts, and time to serve as a single executive KPI. Origin is often attributed to Jack Smyth at [[entity-jellyfish-d6]].


## Related across articles
- [[concept-brand-code]]


## Related across segments
- [[concept-share-of-model-d10]]
- [[concept-share-of-model-d25]]
- [[concept-ai-recall-share]]
- [[concept-mention-rate]]