> ## Documentation Index
> Fetch the complete documentation index at: https://docs.shareofmodel.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Recommendations & Briefs
> How Share Of Model turns measurement into marketing action — recommendations, briefs and execution.
A **recommendation** is the bridge between measurement and execution. It turns a measurable gap surfaced by an analysis — a perception weakness, a visibility deficit, a source you don't yet earn — into a concrete, scored action a marketing team can run.
Recommendations are the foundation of the [Multi-Channel Brief](/platform/roadmap/coming-soon-multi-channel-opportunities-brief) experience.
## What a recommendation is
A recommendation is a single proposed action attached to:
* a **scope** — which brand, category and country it concerns,
* an **analytical reasoning** — which gap, metric, and competitor justify it,
* a **channel** — where the action should be activated (PDP, YouTube, Reddit, PR, TikTok, Display, ChatGPT/Rufus Ads),
* a **priority score** — Effort, Impact and Criticality, used to rank the backlog,
* an **assigned team** — Content, SEO, Paid Media, PR, etc.,
* a **target market** — the geography or audience the action should target.
Recommendations are **deterministic and analytical**. The same gap always produces the same category of recommendation. Reproducible, auditable, and not subject to model hallucinations.
## Lifecycle
A recommendation moves through an explicit set of states:
The recommendation engine inspects the most recent analysis data and surfaces gaps that justify action — brand-mention gaps vs. competitors, weak domain authority in AI citations, under-performing channels, declining visibility trends.
A user reviews the recommendation. Three actions are available:
* **Validate** — accept and trigger brief generation.
* **Reject** — dismiss; the recommendation won't reappear.
* **Hold** — defer for later review.
Once validated, a production-ready brief is generated automatically, no extra scoping required.
The brief is handed off — to your agency, an automated Pencil agent, or executed directly inside Pencil.
## Anatomy of a brief
Every brief shares the same structure so handoffs stay predictable:
| Section | Content |
| ------------------------- | ---------------------------------------------------------------------------- |
| **Action** | Precise description of the task to complete. |
| **Reasoning** | Why this action, why now — backed by Share Of Model data. |
| **Thematic briefing** | Strategic context — the category, the GEO conviction, the competitive shape. |
| **Team & timeline** | Who does what, with what level of urgency. |
| **Production guidelines** | Format, angle, keywords and configuration — ready to execute. |
Briefs are exportable, shareable with agencies, or pushable to Pencil for automated execution.
## How recommendations connect to the pillars
A recommendation is grounded in measurement — it points at one of the [three pillars](/concepts/visibility-perception-influence):
* **Visibility gaps** — you are absent or weak on prompts where competitors dominate.
* **Perception gaps** — your brand is described differently than you'd want, or attributes scored lower than competitors.
* **Influence gaps** — competitors get cited from sources that never reference you.
This anchoring is what makes recommendations defensible: every action has a metric, a competitor and a gap behind it.
## What it is not
Recommendations are **not** generative content. The engine surfaces *what* to do and *why* — using the underlying analytics — but creative production happens through your agency, an automated Pencil agent, or directly inside Pencil. The brief is the contract between insight and execution.
## What's next
The three pillars recommendations build on.
The product surface that exposes recommendations.