# How to Build a Brand Knowledge Base: The D01-D06 Framework

> AI content is only as reliable as its source data. Starbox AI organizes brand facts into six structured libraries to keep output accurate and consistent.

Source: https://www.starbox.work/en/insights/ai-marketing/brand-knowledge-base-d01-d06/ (Starbox AI Insights, updated 2026-10-17)

## Why Build the Knowledge Base First

When you ask AI to write brand content without guardrails, standard problems emerge: it invents selling points, lists wrong prices, adopts competitor terminology, and switches tone from one post to the next. The reason is simple: the AI does not know your brand, so it guesses.

At Starbox AI, we solve this by organizing brand information into a structured knowledge base before any writing begins. Our drafting agents are permitted to pull brand facts exclusively from these files. That is how the output stays reliable.

## The Six Document Libraries

### D01: Product Ammunition

An exhaustive breakdown of every product feature and selling point, mapped directly to the pain point it resolves. Each pain point is translated into experiential descriptions that readers can immediately recognize and feel. Every item is marked as either mandatory to mention or optional based on context.

### D02: Target Audience and Scenario Library

This is the operational core of the entire knowledge base, structured into four chapters:

1. **Decision Type Pool**: What type of purchase decision the customer is making.
2. **Persona Variable Pool**: Customer traits organized by dimension, explicitly identifying which trait combinations are incompatible.
3. **Psychological Scenario Pool**: The exact situations and moments that trigger customer demand.
4. **Emotion and Response**: The customer's emotional state, the strategic direction of the brand's response, and the corresponding writing framework.

### D03: Writing Rules

Hard rules (such as prohibiting fake profound contrasts, forbidding fabricated details, and banning artificial consumer roleplay), brand persona guidelines, stylistic voice attributes, and general editorial standards.

### D04: Brand-Specific Constraints

Strictly banned words, competitor lists, authorized pricing language, category boundaries, and prohibited emotional tones.

### D05: IP and Story Library

Verified background stories about the company and its founders, established whenever founder-led or personal IP content is required.

### D06: Proprietary Enterprise Knowledge Base

Operational business knowledge, service delivery workflows, proprietary methodologies, technical product specifications, and broader industry context.

## Why Selling Points Do Not Drive Topic Selection

A common mistake in marketing is letting product selling points dictate editorial topics. When that happens, every piece of content turns into an aggressive sales pitch. 

Our rule at Starbox AI is simple: **topics must originate from target audiences and their real-life scenarios, while selling points serve purely as ammunition during the drafting stage.** Address the reader's practical problem first, and introduce the product features naturally where they belong.

## Switching Brands Only Requires Swapping the D Layer

The Starbox Content Factory runs on three distinct operational layers:
* Universal rule libraries (fact-extraction rules, headline formulas, advertising compliance, and visual guidelines)
* The brand knowledge base (D01 through D06)
* Style sample libraries

The universal rules, style samples, and underlying AI agents are reused across accounts. When onboarding a new client or product line, our team only needs to populate the D layer.