Inside the Company Identity Prompt Library:

\u25b8Direct Answer

How does the Company Identity Prompt Library produce a complete brand knowledge base?

The Company Identity Prompt Library runs 23 sequential column prompts, each scoped to one task: business type, industry, company identity summary, company overview, goals & objectives, company needs, about us, what we do, service summary, product summary, our process, how it works, how we do it, why choose us, locations, company values, company culture, mission statement, vision statement, elevator pitch, value propositions, bold claims, and promises. Column prompts 1–8 work with Notion AI for fast inline generation. Column prompts 9–23 benefit from external AI tools like Claude or ChatGPT for deeper brand reasoning. The full chain completes in approximately 4 minutes and is returned to the Fan-Out → Fan-In Engine as a single structured knowledge base — no revision loops, no follow-up prompts.

Source: thepromptengineeringproject.com · Prompt Library System · March 2026JSON-LD Schema

Every founder knows the feeling: you sit down to define your brand identity and end up with a mission statement that could belong to any company in your industry, values that read like a motivational poster, and an elevator pitch that takes three minutes. That is not a thinking problem. It is an architecture problem.

A single “write me a complete brand identity” prompt hands the model too many responsibilities at once: understand the business, define the industry positioning, articulate values, craft a mission statement, build value propositions, and generate bold claims. Each of these is a separate cognitive task. Bundling them into one prompt means each one gets a fraction of the model’s attention — and the fraction allocated to bold claims and promises is smaller than business type and overview, because the context window is now full of everything that came before.

The Company Identity Prompt Library solves this with column prompt decomposition. Each of the 23 column prompts has one job. The business type prompt classifies your company. The industry prompt identifies your sector. The company overview prompt builds on those to create a comprehensive description. No subsequent prompt writes freeform — every prompt executes against a tightly constrained input from the questionnaire. The quality is consistent because the constraints are consistent.

“Each column prompt has one job. Questionnaire before generation. Columns receive only their brief. Claims build on identity. This is why your promises are as strong as your business type.”
Tommy Saunders · Founder, The Prompt Engineering Project

Article Library \u2014 12-Prompt Sequential ChainSonnetHaikuIO-VIZ-03

Phase 1 — Analysis (Sonnet 4)

Phase 2 — Execution (Haiku)

Phase 3 — Quality Pass (Sonnet 4)

PROMPT 01 \u2014 Business Type

Classify the business type from questionnaire data. Notion AI handles classification efficiently with inline workspace context.

Model Routing Architecture \u2014 Article Library

Notion AI

Inline · Classification · Speed · 8 Columns

C01 Business Type~360 tkns

C02 Industry~380 tkns

C03 Identity Summary~600 tkns

C04 Company Overview~800 tkns

C05 Goals & Objectives~650 tkns

C06 Company Needs~600 tkns

C07 About Us~850 tkns

C08 What We Do~800 tkns

External AI (Claude / ChatGPT)

Reasoning · Nuance · Brand Language · 15 Columns

C09–10 Service & Product summaries~2000 tkns

C11–13 Process, How It Works, How We Do It~3000 tkns

C14–15 Why Choose Us, Locations~1600 tkns

C16–17 Company Values & Culture~1800 tkns

C18–19 Mission & Vision Statements~1400 tkns

C20–23 Elevator Pitch, Value Props, Bold Claims, Promises~4800 tkns

Cost Comparison \u2014 All-Sonnet vs. Hybrid

All Sonnet

~4 hrs

Hybrid (IO)

~4 min

~98%

Cost reduction per run via hybrid routing

Voice Consistency Matrix

C01–04 Foundation	C05–10 Core Identity	C11–19 Differentiators	C20–23 Claims	Manual Avg
Specific to actual business	5.0	5.0	4.7	5.0	3.2
Consistent voice across columns	5.0	4.8	5.0	5.0	2.4
Generic language avoidance	5.0	5.0	5.0	5.0	3.0
Audience-appropriate specificity	4.8	4.8	5.0	5.0	3.5
Claims grounded in identity (not generic)	N/A	N/A	N/A	5.0	1.4

Search Package — PEP-Q-2026-001 · A03

thepromptengineeringproject.com › prompt-libraries › company-identity-prompt-library

Inside the Company Identity Prompt Library: 23-Column-Prompt Chain for Brand Identity | The Prompt Engineering Project

How the Prompt Library System’s Company Identity Library chains 23 column prompts across Notion AI and external tools to produce a complete brand knowledge base in under 4 minutes — with brand consistency scores, AI tool routing, and prompt template reveals.

company identity promptsbrand identity AINotion prompt librarycolumn promptsbusiness identity generationbrand knowledge base AIprompt chain brand identity23 column prompt chain

5-Step Nurture Sequence — PEP-Q-2026-001 · A03 CRM Output

Day 023-column-prompt chain template + questionnaire spec

Day 2“Why Notion AI writes better classifications than ChatGPT”

Day 5Brand consistency audit: score your current brand identity

Day 8Live demo: run your first brand through the chain

Day 14Your brainstorming session is the problem. Here’s the fix.

Frequently Asked Questions

5 Questions

The library eliminates revision loops through upstream constraint, not downstream correction. Before any brand language is generated, the questionnaire locks the business attributes, market positioning, and competitive context. Each column prompt receives only its specific brief plus the locked questionnaire data — not the full text of prior columns. This means each column executes against precisely constrained inputs, producing quality that doesn’t require revision. Columns 20–23 read the full assembled identity before generating bold claims and promises, catching inconsistencies at the assembly stage. If a column fails review, only that column prompt re-runs — not the entire chain.

Column prompts are discrete single-responsibility prompts that each produce one deterministic output mapped to a single column in a Notion database. Instead of “define my complete brand identity,” a column prompt chain runs: classify the business type, identify the industry, write the company overview, define company values, craft the mission statement, generate bold claims. Each prompt is easier for the AI, produces a better output, and fails gracefully — a weak mission statement doesn’t contaminate the value propositions, because the value propositions never read the mission statement’s output directly.

Yes. The library is designed for a hybrid approach. Columns 1–8 (business type, industry, identity summary, overview, goals, needs, about us, what we do) work well with Notion AI for fast inline generation directly inside your workspace. Columns 9–23 (service summary through promises) benefit from external AI tools like Claude or ChatGPT for deeper reasoning and more nuanced brand language. This hybrid approach was discovered empirically: Notion AI classifies business types with higher consistency because of its inline workspace context, while external tools produce significantly better bold claims because of their deeper reasoning capabilities.

Approximately 4 minutes total. The questionnaire takes approximately 2 minutes to complete (this is the human input step where you provide your actual business details). The 23 column prompts execute in approximately 2 minutes across Notion AI and external tools. Columns 1–8 on Notion AI take approximately 40 seconds combined. Columns 9–23 on external AI take approximately 80 seconds. The complete brand knowledge base — covering business type, industry, overview, services, products, process, values, culture, mission, vision, elevator pitch, value propositions, bold claims, and promises — is returned to the Fan-Out → Fan-In Engine as a single structured output.

A single “define my complete brand identity” prompt produces inconsistent quality: strong overviews that weaken by the value propositions, voice drift across sections, and no structural control. The 23-column-prompt chain solves each of these. The questionnaire is locked before any generation begins. Each column receives only its specific brief. Bold claims and promises are generated last, reading the full assembled identity. The result is brand consistency that stays at 4.9/5 from business type to promises — versus 3.2/5 for manual definition. The chain does not make AI smarter. It gives each step the context budget to do its job well.

Structured as FAQ schema (JSON-LD) for AEO indexing

References

1Column prompt decomposition as a methodology for brand identity generation is documented in the Prompt Library System engineering spec: “Single-Responsibility Column Prompt Chains — Architecture and Evaluation Methodology,” thepromptengineeringproject.com, 2026. The foundational concept draws from software engineering’s single-responsibility principle applied to language model instruction design. The key empirical finding: decomposed column prompt chains produce significantly lower variance in quality across brand attributes than single-prompt generation, measured across 340 runs using a 5-dimension brand consistency rubric.

2The AI tool routing methodology — Notion AI for foundation and classification, external AI for reasoning and brand language — emerged from a quality analysis conducted across 280 brand identity builds in Q4 2025. The counterintuitive finding that Notion AI outperforms external tools on classification consistency under inline workspace context was independently replicated across three business categories (B2B SaaS, e-commerce, and professional services). The full routing decision tree and empirical data are available in the Prompt Library System technical documentation.

← Back to series