Prompt Library ArchitectureDeep Dives

Inside the Company Identity Prompt Library: How 23 Prompts Build Your Brand DNA

Business type to bold claims. A complete brand knowledge base without a single revision loop.

The Prompt Engineering Project March 29, 2026 10 min read

Most companies cannot articulate their own identity clearly. Ask a founder for their mission statement and you get a different answer on Tuesday than you got on Monday. Ask the marketing team for the brand values and they pull up a slide deck from 2022 that nobody has looked at since the offsite where it was created. Ask the sales team for the competitive positioning and they improvise something plausible. The identity exists in fragments -- scattered across pitch decks, About pages, investor memos, and the heads of people who were in the room when the decisions were made.

The Company Identity Prompt Library solves this by generating a complete, coherent brand identity from a single questionnaire input. Twenty-three column prompts chain from business type to bold claims, producing mission, vision, values, positioning, competitive advantages, differentiators, and brand narrative -- all in under four minutes, all from one source of truth, and all without a single revision loop.

This article walks through the twenty-three prompt chain, the six identity pillars it produces, how prompts chain together within the library, and how the Notion implementation makes the output immediately usable.

The Identity Problem

Brand identity is not a creative exercise. It is a structural one. The mission must align with the vision. The values must support the mission. The positioning must reflect the values. The competitive advantages must substantiate the positioning. The differentiators must be provable, not aspirational. And the bold claims -- the statements that a brand uses to command attention -- must be grounded in everything that came before them.

When humans build brand identity manually, they typically work top-down: start with the mission, derive the values, write the positioning. This works in theory but fails in practice because each step is a separate creative session, often with different people, separated by days or weeks. By the time the team reaches competitive advantages, they have forgotten the exact phrasing of the mission. The positioning uses language that contradicts the values. The bold claims sound exciting but have no connection to the differentiators.

The alternative -- asking one prompt to generate everything at once -- fails for a different reason. A single prompt tasked with producing a mission statement, vision, values, positioning, competitive advantages, differentiators, and bold claims simultaneously cannot give adequate attention to any of them. The outputs are shallow, generic, and interchangeable between companies. They read like they were generated by a model -- because they were, by a model that was asked to do too many things at once.

Prompts in the Chain

Identity Pillars

Minutes to Complete

Revision Loops

The 23-Prompt Chain

The Company Identity library contains twenty-three column prompts organized into six pillars. Each prompt produces one specific output. The prompts within each pillar are designed to be independent -- they all read from the questionnaire, not from each other's outputs. This means they can execute in parallel within their pillar, and pillars can execute in parallel with each other.

The chain is not strictly sequential. It is better described as a directed acyclic graph with three layers. Layer one: foundation prompts that read only from the questionnaire. Layer two: synthesis prompts that read from the questionnaire and from layer one outputs. Layer three: expression prompts that read from layer two outputs to produce the final, polished identity artifacts.

Company Identity Prompt Chain

Layer 1 — Foundation (parallel, questionnaire-only)
├── Business Type Analysis
├── Market Position Assessment
├── Audience Need Mapping
├── Competitive Landscape Summary
├── Core Capability Inventory
├── Value Proposition Extraction
└── Brand Personality Synthesis

Layer 2 — Synthesis (parallel, reads Layer 1 + questionnaire)
├── Mission Statement Generation
├── Vision Statement Generation
├── Core Values Derivation
├── Strategic Positioning
├── Competitive Advantage Mapping
├── Key Differentiator Identification
├── Brand Narrative Arc
└── Unique Selling Propositions

Layer 3 — Expression (parallel, reads Layer 2 + questionnaire)
├── Mission Statement (polished)
├── Elevator Pitch (30-second)
├── Elevator Pitch (60-second)
├── Bold Claims Generation
├── Brand Promise Statement
├── Tagline Candidates
├── Brand Manifesto
└── Identity Summary Record

The three-layer structure is important because it mirrors how identity actually works. You cannot write a positioning statement without first understanding the market position and audience needs (Layer 1). You cannot write bold claims without first having a positioning statement and competitive advantages (Layer 2). The layers enforce logical dependency without requiring serial execution -- everything within each layer runs in parallel.

The total execution time is determined by the slowest prompt in each layer, not by the sum of all prompts. Layer 1 typically completes in 45 seconds. Layer 2 in 60 seconds. Layer 3 in 90 seconds. Total: under four minutes, including validation and assembly.

The Six Identity Pillars

The twenty-three prompts produce outputs that organize into six identity pillars. Each pillar represents a distinct facet of brand identity, and together they form a complete identity system.

Pillar 1: Mission and Vision

The Mission and Vision pillar produces three artifacts: a mission statement (what the company does and why), a vision statement (what the world looks like when the company succeeds), and a mission-vision alignment analysis that verifies the two statements are complementary rather than contradictory. The mission prompt reads business type, audience needs, and core capabilities from the questionnaire. The vision prompt reads the mission output plus the company's stated aspirations. The alignment analysis reads both and flags any tension.

Pillar 2: Values

The Values pillar produces a set of three to five core values, each with a definition, a behavioral description (what living this value looks like in practice), and an anti-pattern (what violating this value looks like). Values are not aspirational words pulled from a list -- they are derived from the questionnaire's brand personality descriptors, origin story, and competitive positioning. A company whose origin story centers on speed and whose competitive advantage is rapid deployment will not get "patience" as a core value.

Pillar 3: Positioning

The Positioning pillar produces the strategic positioning statement (the classic "for [audience] who [need], [company] is the [category] that [benefit]" framework), the unique selling propositions, and the market category definition. Positioning is where the competitive context from the questionnaire becomes critical -- the system cannot position a company in a market it does not understand. The positioning prompt reads the full competitive landscape summary from Layer 1 and the value proposition extraction.

Pillar 4: Differentiators

The Differentiators pillar identifies what makes the company genuinely different from competitors -- not different in the "we also do X but better" sense, but different in the "competitors cannot easily replicate this" sense. The differentiator prompts cross-reference the competitive landscape with the company's core capabilities and origin story. A differentiator must be substantiated -- the prompt rejects claims it cannot ground in questionnaire data.

Pillar 5: Competitive Advantages

The Competitive Advantages pillar maps specific advantages against specific competitors. Unlike differentiators (which are absolute), competitive advantages are relative -- "we are faster than Competitor A" and "we are more affordable than Competitor B." The output is structured as a matrix: competitors on one axis, advantage categories on the other, with evidence from the questionnaire in each cell. This matrix feeds directly into the sales enablement library's battle card generation.

Pillar 6: Bold Claims

The Bold Claims pillar produces the attention-commanding statements that a brand uses in headlines, conference talks, and social media. Bold claims are the riskiest output because they must be provocative enough to capture attention but grounded enough to be defensible. The bold claims prompt reads the full identity stack -- mission, vision, values, positioning, differentiators, and competitive advantages -- and generates claims that are substantiated by at least two other pillars.

A bold claim without a foundation is marketing fiction. Every bold claim generated by the system traces back to at least two identity pillars.

How Prompts Chain Together

The chaining mechanism is not sequential prompt feeding -- it is structured context injection. When a Layer 2 prompt needs the output of a Layer 1 prompt, it receives that output as a formatted context block within its input, separate from the questionnaire data. The prompt sees two distinct inputs: "here is the questionnaire data" and "here is the output from the foundation analysis." This separation is critical because it prevents the model from confusing source data (questionnaire) with derived data (Layer 1 outputs).

context-injection.ts

// Layer 2 prompt receives structured context
const layer2Input = {
  // Original questionnaire data (source of truth)
  questionnaire: {
    brand_identity: { ... },
    key_message: { ... },
    competitive_context: { ... },
  },

  // Layer 1 outputs (derived data, clearly labeled)
  foundation: {
    business_type_analysis: "...",    // From Layer 1
    market_position: "...",           // From Layer 1
    core_capabilities: ["..."],       // From Layer 1
    value_proposition: "...",         // From Layer 1
  },

  // Instruction for this specific column
  instruction: "Generate the strategic positioning statement...",
};

The context injection approach has a key advantage over sequential chaining: the context window does not grow between prompts. Each prompt receives only its required inputs -- the questionnaire fields it needs and the specific Layer 1 outputs it depends on. A Layer 2 prompt that needs three Layer 1 outputs receives those three outputs plus its questionnaire fields, not the entire Layer 1 output set. The context window stays flat regardless of how many prompts have executed.

Error handling at the chain level is straightforward. If a Layer 1 prompt fails, only the Layer 2 prompts that depend on it are blocked. Other Layer 2 prompts that do not require the failed output continue executing normally. The system reports partial completion -- "18 of 23 columns populated, 5 columns blocked by foundation analysis failure" -- rather than failing entirely.

Monitor Layer 1 prompt quality aggressively. A weak foundation analysis propagates through Layer 2 and Layer 3. Catching a quality issue at Layer 1 is ten times cheaper than catching it at Layer 3, because you only re-run one prompt instead of the entire chain.

Notion Implementation

The Company Identity library is implemented as a Notion database where each row represents one company's identity and each column maps to one column prompt's output. The database has twenty-three content columns (one per prompt), plus metadata columns for the company name, questionnaire version, generation date, and quality status.

Notion AI plays a specific role in the implementation. For prompts that require synthesis across multiple fields, Notion AI processes the questionnaire data within the Notion environment, leveraging its native understanding of the database schema. For prompts that require deeper reasoning or larger context windows, the system dispatches to external AI through API calls and writes the results back to the Notion database.

Notion-native prompts for simple extraction

Prompts that primarily reformulate questionnaire data -- like extracting brand personality descriptors into a structured format -- run natively in Notion AI, minimizing latency and API costs.

External AI for complex synthesis

Prompts requiring multi-step reasoning -- like competitive advantage mapping or bold claims generation -- dispatch to Claude or GPT-4 through the API, with results written back to Notion columns.

Hybrid execution for layered prompts

Layer 1 prompts run primarily in Notion. Layer 2 and Layer 3 prompts use external AI with Notion-formatted context. This keeps the simple work fast and cheap while reserving API costs for the prompts that need them.

Views for different stakeholders

Founders see Mission, Vision, Values, and Bold Claims. Marketing sees Positioning, Differentiators, and Brand Narrative. Sales sees Competitive Advantages and USPs. Each view is a filtered, sorted perspective on the same database.

The Notion implementation also enables the zero-revision-loop claim. Because every output is visible in the database immediately after generation, stakeholders can review the full identity simultaneously. If the mission statement needs adjustment, the user refines the relevant questionnaire fields and re-runs only the mission-related column prompts. There is no "send it back for revision" loop because there is no intermediary -- the output is in the database, and the input is in the questionnaire. Adjust one, re-run, and the other updates.

The Company Identity Prompt Library demonstrates what happens when you apply engineering discipline to a creative problem. Brand identity is not less creative because it is generated by twenty-three structured prompts -- it is more coherent, more substantiated, and more useful. Every pillar connects to every other pillar. Every bold claim traces back to a competitive advantage. Every value connects to the mission. The structure does not constrain the output -- it ensures the output holds together as a system.

The zero-revision-loop is not a gimmick. It is a consequence of the architecture. When every output traces to a structured input, revision means adjusting the input and re-running the relevant prompts. There is no creative director who needs to approve the seventh draft. There is a questionnaire, a prompt library, and a knowledge base. Adjust, run, review. The loop is one step, and it takes four minutes.

Key Takeaways

The Company Identity library uses 23 column prompts organized in three layers: foundation (questionnaire analysis), synthesis (identity generation), and expression (polished artifacts).

Six identity pillars -- Mission/Vision, Values, Positioning, Differentiators, Competitive Advantages, and Bold Claims -- form a complete, interconnected identity system.

Context injection replaces sequential chaining. Each prompt receives its specific dependencies as structured context blocks, keeping the context window flat regardless of chain depth.

Notion implementation splits execution between Notion AI (simple extraction) and external AI (complex synthesis), with database views tailored to different stakeholders.

Zero revision loops result from the architecture: adjust the questionnaire input, re-run affected column prompts, and the knowledge base updates in minutes without creative review cycles.

The Context Brief: The One Document That Runs Your Entire Stack Inside the Article Library: How the Writing Engine Produces Long-Form at Scale

Prompt Library Architecture