Demo: Yowie Marketing Ecosystem

A DTC outdoor gear brand with four product lines and a 6-person team. An interconnected content pipeline built with Claude Code — from brand configuration through research, content generation, and human review. Scroll through to see how.

Yowie is an outdoor gear company with a 6-person team managing four product lines across four audience segments. This project demonstrates how Claude Code can build an interconnected ecosystem of marketing tools, all governed by a single brand configuration — with human decision gates between every AI stage.

Each tool serves a unique purpose but they all share the same data and inform each other. See the GitHub repo for brand, data, and related context.

Architecture Map

The full pipeline: Every tool reads from the Brand Configuration Layer. Every AI output passes through a human decision gate before moving downstream. The pipeline flows from research through content generation to final delivery across email, social, CMS, and design.

tools/architecture_map.html Open Full Size →

Brand Configuration Viewer

Problem it solves: Brand voice lives in a PDF that nobody reads. Copy sounds different depending on who wrote it, which channel it's for, and whether the writer remembered to check the style guide. The result is inconsistent messaging across products and audiences — backcountry hikers get casual copy, car campers get jargon they don't understand.

Prompt: Build an interactive brand configuration viewer. Read the brand definition and output a structured JSON configuration that captures voice rules, channel-specific registers, audience adaptations, messaging pillars with audience resonance, approved/forbidden/conditional terminology, and 22 hand-crafted copy examples across all product/audience/channel combinations. Build an interactive HTML viewer where I can select any product, audience, and channel and see the applicable rules, pillar rankings, terminology, and example copy.

tools/brand_config_viewer.html Open Full Size →

Research & Brief Builder

Problem it solves: Campaign briefs are written from scratch every time, with competitive research scattered across browser tabs and audience insights trapped in someone's head. The brief that reaches the content creator is usually a Slack message: "write something for the cooler launch, make it sound good." No structured inputs, no reasoning trail, no way to reproduce what worked.

Prompt: Build the Research & Brief Builder. It takes raw inputs (product data, audience research, competitive landscape, seasonal context) and produces structured campaign briefs that feed directly into the Content Engine. Demonstrate it by building a brief for the cooler market acquisition campaign: research the competitive cooler landscape (YETI, Stanley, RTIC, Igloo), analyze the casual buyer audience, identify Yowie's positioning angle, and output a complete campaign brief. Include a pipeline diagram showing human decision points in the workflow.

tools/brief_builder.html Open Full Size →

Content Engine

Problem it solves: Producing content for 8 products across 4 audience segments and 4 channels means 128 potential content variations. A 6-person team can't write them all, and generic AI copy doesn't know that backcountry hikers want raw specs while car campers want outcomes translated into plain English. The Content Engine reads the brand configuration and an approved campaign brief, then produces content that's on-brand, segment-specific, and channel-ready.

Prompt: Build the Content Engine. It reads the Brand Configuration Layer and takes a structured campaign brief as input. The engine produces on-brand, segment-specific, channel-ready content. Demonstrate four output modes: single item, segment variants (same message adapted for backcountry vs. casual buyers), channel variants (email, social, landing page from one brief), and a full campaign package. Show rules applied, pillar used, and rationale for every piece.

tools/content_engine.html Open Full Size →

Build tools like these for your team.

Agency State teaches professionals how to capture collective knowledge and put AI to work with it.

Visit Agency State